fix(handler): make on_comment and @mention paths mutually exclusive

Without (issue, agent) coalescing, a single member comment that @mentions
the assignee was enqueueing two tasks with identical trigger_comment_id —
once via shouldEnqueueOnComment → EnqueueTaskForIssue and once via
enqueueMentionedAgentTasks → EnqueueTaskForMention. The same double-fire
hit plain replies that inherit the assignee mention from the thread root.

Add commentMentionsAssignee, which uses the shared
shouldInheritParentMentions logic to compute the same effective mention
set the @mention path will see. Extend the on_comment gate to skip when
that helper says the assignee will be triggered through the @mention
path. Net result: exactly one task per (comment, assignee), with the
trigger comment preserved.

TestOnCommentTriggerDecision updated to reflect the new contract; assignee
mention cases that used to assert the on_comment branch fires now assert
it skips (mention path covers them). New integration test
TestAssigneeMentionDoesNotDoubleEnqueue pins the end-to-end behavior.

Also softens the "ClaimAgentTask guarantees serial execution" wording in
the code and the RFC: it is a coordination-side property under the
current single-poller-per-runtime model, not a DB-level lock on
(issue, agent). FOR UPDATE SKIP LOCKED locks the row being claimed, not
the key.

Co-authored-by: multica-agent <github@multica.ai>

.dockerignore

@@ -0,0 +1,38 @@
+# Dependencies
+node_modules
+.pnpm-store
+
+# Build outputs
+.next
+dist
+server/bin
+server/tmp
+
+# Git
+.git
+.gitignore
+
+# Environment
+.env
+.env.*
+!.env.example
+
+# IDE
+.idea
+.vscode
+*.swp
+*.swo
+
+# OS
+.DS_Store
+Thumbs.db
+
+# Test
+e2e/test-results
+coverage
+
+# Docs
+docs/
+
+# Desktop app (not needed for web self-hosting)
+apps/desktop

.env.example

@@ -4,9 +4,28 @@ POSTGRES_USER=multica
 POSTGRES_PASSWORD=multica
 POSTGRES_PORT=5432
 DATABASE_URL=postgres://multica:multica@localhost:5432/multica?sslmode=disable
+# Optional pgxpool tuning. Defaults are 25 / 5 per pod and are usually fine.
+# You can also set pool_max_conns / pool_min_conns as query params on
+# DATABASE_URL; env vars below take precedence over URL params.
+# DATABASE_MAX_CONNS=25
+# DATABASE_MIN_CONNS=5
 
 # Server
+# APP_ENV gates production safety checks. Docker self-host pins APP_ENV to
+# "production" by default. Local dev can leave it unset.
+# See SELF_HOSTING.md for the full login setup.
+APP_ENV=
+# Optional local/testing shortcut. Empty by default, so there is no fixed
+# verification code. Without RESEND_API_KEY, generated codes print to stdout.
+# If you need deterministic local automation, set a 6-digit value such as
+# 888888 and keep APP_ENV non-production. This is ignored when APP_ENV=production.
+MULTICA_DEV_VERIFICATION_CODE=
 PORT=8080
+# Prometheus metrics are disabled by default. When enabled, bind to loopback
+# unless you protect the listener with private networking, allowlists, or
+# proxy auth. Do not expose this endpoint through the public app/API ingress.
+# HTTP request metrics start accumulating only when this listener is enabled.
+# METRICS_ADDR=127.0.0.1:9090
 JWT_SECRET=change-me-in-production
 MULTICA_SERVER_URL=ws://localhost:8080/ws
 MULTICA_APP_URL=http://localhost:3000
@@ -21,30 +40,100 @@ MULTICA_CODEX_MODEL=
 MULTICA_CODEX_WORKDIR=
 MULTICA_CODEX_TIMEOUT=20m
 
+# Self-host image channel
+# Default stable release channel. Pin to an exact release like v0.2.4 if you
+# want to stay on a specific version. If the selected tag has not been
+# published to GHCR yet, use make selfhost-build / the build override instead.
+MULTICA_IMAGE_TAG=latest
+MULTICA_BACKEND_IMAGE=ghcr.io/multica-ai/multica-backend
+MULTICA_WEB_IMAGE=ghcr.io/multica-ai/multica-web
+
 # Email (Resend)
+# For local/dev use, leave RESEND_API_KEY empty — generated codes print to stdout.
+# For production, set your Resend API key and change RESEND_FROM_EMAIL to a domain verified in your Resend account.
 RESEND_API_KEY=
 RESEND_FROM_EMAIL=noreply@multica.ai
 
 # Google OAuth
+# The web login page reads GOOGLE_CLIENT_ID from /api/config at runtime, so
+# changing it only requires restarting the backend / compose stack. No web
+# rebuild is needed.
 GOOGLE_CLIENT_ID=
 GOOGLE_CLIENT_SECRET=
 GOOGLE_REDIRECT_URI=http://localhost:3000/auth/callback
 
 # S3 / CloudFront
+# S3_BUCKET — bucket NAME only (e.g. "my-bucket"). Do NOT include the
+# ".s3.<region>.amazonaws.com" suffix; the server builds the public URL
+# from S3_BUCKET + S3_REGION. S3_REGION must match the bucket's real region.
 S3_BUCKET=
 S3_REGION=us-west-2
 CLOUDFRONT_KEY_PAIR_ID=
 CLOUDFRONT_PRIVATE_KEY_SECRET=multica/cloudfront-signing-key
 CLOUDFRONT_PRIVATE_KEY=
 CLOUDFRONT_DOMAIN=
+# COOKIE_DOMAIN — optional Domain attribute on session + CloudFront cookies.
+# Leave empty for single-host deployments (localhost, LAN IP, or a single
+# hostname) — session cookies become host-only, which is what the browser
+# wants. Only set it when the frontend and backend sit on different
+# subdomains of one registered domain (e.g. ".example.com"). Do NOT set it
+# to an IP address: RFC 6265 forbids IP literals in the cookie Domain
+# attribute and browsers silently drop such cookies.
 COOKIE_DOMAIN=
 
+# Local file storage (fallback when S3_BUCKET is not set)
+LOCAL_UPLOAD_DIR=./data/uploads
+LOCAL_UPLOAD_BASE_URL=http://localhost:8080
+
+# Security
+# Comma-separated list of allowed origins for CORS and WebSocket connections.
+# Defaults to localhost dev origins when unset.
+# Example: ALLOWED_ORIGINS=https://app.multica.ai,https://staging.multica.ai
+ALLOWED_ORIGINS=
+
+# Realtime metrics endpoint (/health/realtime) access control. See MUL-1342.
+# When unset, the endpoint only serves direct loopback (127.0.0.1 / ::1)
+# callers with no forwarding headers and returns 404 to everything else —
+# safe for local dev. Any deployment behind a reverse proxy (Caddy / Nginx
+# terminating TLS in front of localhost:8080) MUST set this token, since
+# proxied requests look like loopback at the Go layer; with no token, those
+# requests are refused with 404. Pass the token as
+# `Authorization: Bearer <token>`.
+# REALTIME_METRICS_TOKEN=
+
 # Frontend
 FRONTEND_PORT=3000
 FRONTEND_ORIGIN=http://localhost:3000
-NEXT_PUBLIC_API_URL=http://localhost:8080
-NEXT_PUBLIC_WS_URL=ws://localhost:8080/ws
+# Leave empty — auto-derived from page origin in browser, set by Makefile for local dev.
+# Only set explicitly if frontend and backend are on different domains.
+NEXT_PUBLIC_API_URL=
+NEXT_PUBLIC_WS_URL=
 
 # Remote API (optional) — set to proxy local frontend to a remote backend
 # Leave empty to use local backend (localhost:8080)
 # REMOTE_API_URL=https://multica-api.copilothub.ai
+
+# ==================== Self-hosting: Control Signups (fixes #930) ====================
+# Set to "false" to completely disable new user signups (recommended for private instances)
+ALLOW_SIGNUP=true
+# The web UI reads ALLOW_SIGNUP from /api/config at runtime, so toggling this
+# only requires restarting the backend / compose stack — not rebuilding web.
+# It is not hot-reloaded.
+
+# Optional: Only allow emails from these domains (comma-separated)
+ALLOWED_EMAIL_DOMAINS=
+
+# Optional: Only allow these exact email addresses (comma-separated)
+ALLOWED_EMAILS=
+
+# ==================== Analytics (PostHog) ====================
+# Product analytics events feed the acquisition → activation → expansion funnel.
+# Leave POSTHOG_API_KEY empty for local dev / self-hosted instances; the server
+# will run a no-op analytics client and ship nothing. See docs/analytics.md.
+POSTHOG_API_KEY=
+POSTHOG_HOST=https://us.i.posthog.com
+# Optional override for the `environment` PostHog event property.
+# Defaults from APP_ENV and normalizes to production / staging / dev.
+ANALYTICS_ENVIRONMENT=
+# Force the no-op client even when POSTHOG_API_KEY is set (CI / opt-out).
+ANALYTICS_DISABLED=

.gitattributes

@@ -0,0 +1,6 @@
+# Ensure shell scripts always use LF line endings (needed for Docker on Windows)
+*.sh text eol=lf
+docker/entrypoint.sh text eol=lf
+
+# Default behavior
+* text=auto

.github/ISSUE_TEMPLATE/bug_report.yml

@@ -0,0 +1,50 @@
+name: "Bug Report"
+description: Report a bug — something that's broken, crashes, or behaves incorrectly.
+title: "[Bug]: "
+labels: ["bug"]
+body:
+  - type: dropdown
+    id: deployment
+    attributes:
+      label: Deployment type
+      description: Are you using the hosted version or a self-hosted instance?
+      options:
+        - multica.ai (hosted)
+        - Self-hosted
+    validations:
+      required: true
+
+  - type: textarea
+    id: description
+    attributes:
+      label: What happened?
+      description: Describe the bug and what you expected instead. Screenshots, error messages, or screen recordings are welcome.
+      placeholder: |
+        When I do X, Y happens. I expected Z instead.
+    validations:
+      required: true
+
+  - type: textarea
+    id: reproduction
+    attributes:
+      label: Steps to reproduce
+      description: How can we trigger this bug?
+      placeholder: |
+        1. Go to '...'
+        2. Click on '...'
+        3. See error
+    validations:
+      required: true
+
+  - type: textarea
+    id: screenshots
+    attributes:
+      label: Screenshots (optional)
+      description: If applicable, add screenshots or screen recordings to help explain the problem.
+
+  - type: textarea
+    id: context
+    attributes:
+      label: Additional context (optional)
+      description: Environment info, logs, or anything else that might help.
+      render: shell

.github/ISSUE_TEMPLATE/config.yml

@@ -0,0 +1 @@
+blank_issues_enabled: true

.github/ISSUE_TEMPLATE/feature_request.yml

@@ -0,0 +1,37 @@
+name: "Feature Request"
+description: Suggest a new feature or improvement.
+title: "[Feature]: "
+labels: ["enhancement"]
+body:
+  - type: dropdown
+    id: deployment
+    attributes:
+      label: Deployment type
+      description: Are you using the hosted version or a self-hosted instance?
+      options:
+        - multica.ai (hosted)
+        - Self-hosted
+    validations:
+      required: true
+
+  - type: textarea
+    id: description
+    attributes:
+      label: What do you want and why?
+      description: Describe the problem you're trying to solve or the improvement you'd like to see.
+      placeholder: |
+        I'm trying to do X but there's no way to...
+    validations:
+      required: true
+
+  - type: textarea
+    id: solution
+    attributes:
+      label: Proposed solution (optional)
+      description: If you have an idea for how this should work, describe it here.
+
+  - type: textarea
+    id: screenshots
+    attributes:
+      label: Screenshots / mockups (optional)
+      description: If applicable, add screenshots, mockups, or sketches to illustrate your idea.

.github/PULL_REQUEST_TEMPLATE.md

@@ -0,0 +1,60 @@
+## What does this PR do?
+
+<!-- Describe the change clearly. What problem does it solve? Why is this approach the right one? -->
+
+
+
+## Related Issue
+
+<!-- Link the issue this PR addresses. If no issue exists, consider creating one first. -->
+
+Closes #
+
+## Type of Change
+
+- [ ] Bug fix (non-breaking change that fixes an issue)
+- [ ] New feature (non-breaking change that adds functionality)
+- [ ] Refactor / code improvement (no behavior change)
+- [ ] Documentation update
+- [ ] Tests (adding or improving test coverage)
+- [ ] CI / infrastructure
+
+## Changes Made
+
+<!-- List the specific changes. Include file paths for code changes. -->
+
+-
+
+## How to Test
+
+<!-- Steps to verify this change works. For bugs: reproduction steps + proof that the fix works. -->
+
+1.
+2.
+3.
+
+## Checklist
+
+- [ ] I have included a thinking path that traces from project context to this change
+- [ ] I have run tests locally and they pass
+- [ ] I have added or updated tests where applicable
+- [ ] If this change affects the UI, I have included before/after screenshots
+- [ ] I have updated relevant documentation to reflect my changes
+- [ ] If I added a new runtime / coding tool / UI tab, I synced the change to **landing copy** (`apps/web/features/landing/i18n/`), **starter-content** (`packages/views/onboarding/utils/starter-content-content-*.ts`), and **relevant docs** (`apps/docs/content/docs/`)
+- [ ] If this PR touches Chinese product copy, I checked it against `apps/docs/content/docs/developers/conventions.zh.mdx` (terminology, mixed-rule for `task` / `issue` / `skill`)
+- [ ] I have considered and documented any risks above
+- [ ] I will address all reviewer comments before requesting merge
+
+## AI Disclosure
+
+<!-- Most PRs involve AI coding tools — that's totally fine! We're curious about your process. -->
+
+**AI tool used:** <!-- e.g. Claude Code, Cursor, GitHub Copilot, Multica Agent, N/A -->
+
+**Prompt / approach:**
+<!-- How did you use AI to produce this code? Share your prompt, conversation link, or describe your approach. This helps the team learn from each other's AI workflows. -->
+
+
+## Screenshots (optional)
+
+<!-- If applicable, add screenshots showing the change in action. -->

.github/workflows/ci.yml

@@ -29,8 +29,17 @@ jobs:
       - name: Install dependencies
         run: pnpm install
 
-      - name: Build, type check, and test
-        run: pnpm build && pnpm typecheck && pnpm test
+      - name: Verify reserved-slugs.ts is up to date
+        # Re-runs the generator and fails on any drift from the
+        # checked-in TypeScript output. The Go side embeds the JSON
+        # source directly, so a passing diff here proves both sides
+        # share one source of truth.
+        run: |
+          pnpm generate:reserved-slugs
+          git diff --exit-code -- packages/core/paths/reserved-slugs.ts
+
+      - name: Build, type check, lint, and test
+        run: pnpm exec turbo build typecheck lint test --filter='!@multica/docs'
 
   backend:
     runs-on: ubuntu-latest
@@ -48,8 +57,22 @@ jobs:
           --health-interval 5s
           --health-timeout 5s
           --health-retries 20
+      redis:
+        image: redis:7-alpine
+        ports:
+          - 6379:6379
+        options: >-
+          --health-cmd "redis-cli ping"
+          --health-interval 5s
+          --health-timeout 5s
+          --health-retries 10
     env:
       DATABASE_URL: postgres://multica:multica@localhost:5432/multica?sslmode=disable
+      # Wires up the RedisLocalSkill*_test.go suite. Distinct from REDIS_URL
+      # (which would flip the server binary itself onto the Redis-backed
+      # realtime relay + request stores); the tests talk to this Redis
+      # directly so they run alongside the Postgres-backed suite.
+      REDIS_TEST_URL: redis://localhost:6379/1
     steps:
       - name: Checkout
         uses: actions/checkout@v6

.github/workflows/desktop-smoke.yml

@@ -0,0 +1,59 @@
+name: Desktop Smoke Build
+
+on:
+  workflow_dispatch:
+
+permissions:
+  contents: read
+
+jobs:
+  desktop:
+    strategy:
+      fail-fast: false
+      matrix:
+        include:
+          - os: ubuntu-latest
+            target: linux
+          - os: windows-latest
+            target: win
+    runs-on: ${{ matrix.os }}
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v4
+        with:
+          fetch-depth: 0
+
+      - name: Install rpmbuild (Linux)
+        if: matrix.target == 'linux'
+        run: sudo apt-get update && sudo apt-get install -y rpm
+
+      - name: Setup Go
+        uses: actions/setup-go@v5
+        with:
+          go-version-file: server/go.mod
+          cache-dependency-path: server/go.sum
+
+      - name: Setup pnpm
+        uses: pnpm/action-setup@v4
+
+      - name: Setup Node
+        uses: actions/setup-node@v4
+        with:
+          node-version: 22
+          cache: pnpm
+
+      - name: Install dependencies
+        run: pnpm install --frozen-lockfile
+
+      - name: Package Desktop installers (${{ matrix.target }})
+        working-directory: apps/desktop
+        env:
+          CSC_IDENTITY_AUTO_DISCOVERY: "false"
+        run: node scripts/package.mjs --${{ matrix.target }} --x64 --arm64 --publish never
+
+      - name: Upload Desktop artifacts (${{ matrix.target }})
+        uses: actions/upload-artifact@v4
+        with:
+          name: desktop-${{ matrix.target }}
+          path: apps/desktop/dist
+          if-no-files-found: error

.github/workflows/release.yml

@@ -3,13 +3,65 @@ name: Release
 on:
   push:
     tags:
-      - "v*"
+      # GitHub Actions uses glob patterns here, not regex. Match versioned
+      # tags broadly at the trigger layer, then enforce strict semver below.
+      - "v*.*.*"
+      - "!v*-dirty*"
 
 permissions:
   contents: write
+  packages: write
 
 jobs:
+  verify:
+    runs-on: ubuntu-latest
+    outputs:
+      tag_name: ${{ steps.release_meta.outputs.tag_name }}
+      is_stable: ${{ steps.release_meta.outputs.is_stable }}
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v4
+        with:
+          fetch-depth: 0
+
+      - name: Validate tag name
+        id: release_meta
+        shell: bash
+        run: |
+          tag="${GITHUB_REF_NAME}"
+          echo "Triggered by tag: $tag"
+          if [[ ! "$tag" =~ ^v[0-9]+\.[0-9]+\.[0-9]+(-[0-9A-Za-z.-]+)?$ ]]; then
+            echo "::error::Release tags must look like vX.Y.Z or vX.Y.Z-suffix; got '$tag'."
+            exit 1
+          fi
+          if [[ "$tag" == *-dirty* ]]; then
+            echo "::error::Refusing to release from dirty tag '$tag'."
+            exit 1
+          fi
+          echo "tag_name=$tag" >> "$GITHUB_OUTPUT"
+          if [[ "$tag" == *-* ]]; then
+            echo "is_stable=false" >> "$GITHUB_OUTPUT"
+          else
+            echo "is_stable=true" >> "$GITHUB_OUTPUT"
+          fi
+
+      - name: Setup Go
+        uses: actions/setup-go@v5
+        with:
+          go-version-file: server/go.mod
+          cache-dependency-path: server/go.sum
+
+      - name: Run tests
+        run: cd server && go test ./...
+
   release:
+    needs: verify
+    # Only run on the canonical upstream repo. Forks don't have the
+    # HOMEBREW_TAP_GITHUB_TOKEN secret and should not be publishing to
+    # `multica-ai/homebrew-tap` anyway. Without this guard, every fork's
+    # tag push fails this job (401 against the upstream tap), which makes
+    # downstream CI go red without affecting the actual artifact pipeline.
+    if: github.repository_owner == 'multica-ai'
     runs-on: ubuntu-latest
     steps:
       - name: Checkout
@@ -23,9 +75,6 @@ jobs:
           go-version-file: server/go.mod
           cache-dependency-path: server/go.sum
 
-      - name: Run tests
-        run: cd server && go test ./...
-
       - name: Run GoReleaser
         uses: goreleaser/goreleaser-action@v6
         with:
@@ -34,3 +83,298 @@ jobs:
         env:
           GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
           HOMEBREW_TAP_GITHUB_TOKEN: ${{ secrets.HOMEBREW_TAP_GITHUB_TOKEN }}
+
+  # Multi-arch images are built natively per platform on dedicated runners
+  # (amd64 on ubuntu-latest, arm64 on ubuntu-24.04-arm) and merged into a
+  # manifest list. This avoids QEMU emulation, which was making the Next.js
+  # arm64 build run for 30+ minutes per release.
+  docker-backend-build:
+    needs: verify
+    strategy:
+      fail-fast: false
+      matrix:
+        include:
+          - platform: linux/amd64
+            runs-on: ubuntu-latest
+          - platform: linux/arm64
+            runs-on: ubuntu-24.04-arm
+    runs-on: ${{ matrix.runs-on }}
+    steps:
+      - name: Prepare
+        run: |
+          platform=${{ matrix.platform }}
+          echo "PLATFORM_PAIR=${platform//\//-}" >> "$GITHUB_ENV"
+
+      - name: Checkout
+        uses: actions/checkout@v4
+
+      - name: Compute backend image labels
+        id: meta
+        uses: docker/metadata-action@v5
+        with:
+          images: ghcr.io/${{ github.repository_owner }}/multica-backend
+          labels: |
+            org.opencontainers.image.title=Multica Backend
+            org.opencontainers.image.description=Multica self-hosted backend
+
+      - name: Setup Docker Buildx
+        uses: docker/setup-buildx-action@v3
+
+      - name: Login to GHCR
+        uses: docker/login-action@v3
+        with:
+          registry: ghcr.io
+          username: ${{ github.actor }}
+          password: ${{ secrets.GITHUB_TOKEN }}
+
+      - name: Build and push by digest
+        id: build
+        uses: docker/build-push-action@v6
+        with:
+          context: .
+          file: Dockerfile
+          pull: true
+          platforms: ${{ matrix.platform }}
+          labels: ${{ steps.meta.outputs.labels }}
+          cache-from: type=gha,scope=release-backend-${{ env.PLATFORM_PAIR }}
+          cache-to: type=gha,mode=max,scope=release-backend-${{ env.PLATFORM_PAIR }}
+          build-args: |
+            VERSION=${{ needs.verify.outputs.tag_name }}
+            COMMIT=${{ github.sha }}
+          outputs: type=image,name=ghcr.io/${{ github.repository_owner }}/multica-backend,push-by-digest=true,name-canonical=true,push=true
+
+      - name: Export digest
+        run: |
+          mkdir -p /tmp/digests
+          digest="${{ steps.build.outputs.digest }}"
+          touch "/tmp/digests/${digest#sha256:}"
+
+      - name: Upload digest
+        uses: actions/upload-artifact@v4
+        with:
+          name: digests-backend-${{ env.PLATFORM_PAIR }}
+          path: /tmp/digests/*
+          if-no-files-found: error
+          retention-days: 1
+
+  docker-backend-merge:
+    needs: [verify, docker-backend-build]
+    runs-on: ubuntu-latest
+    concurrency:
+      group: release-docker-backend-${{ github.ref }}
+      cancel-in-progress: true
+    steps:
+      - name: Download digests
+        uses: actions/download-artifact@v4
+        with:
+          path: /tmp/digests
+          pattern: digests-backend-*
+          merge-multiple: true
+
+      - name: Setup Docker Buildx
+        uses: docker/setup-buildx-action@v3
+
+      - name: Compute backend image tags
+        id: meta
+        uses: docker/metadata-action@v5
+        with:
+          images: ghcr.io/${{ github.repository_owner }}/multica-backend
+          flavor: |
+            latest=false
+          tags: |
+            type=raw,value=latest,enable=${{ needs.verify.outputs.is_stable == 'true' }}
+            type=raw,value=${{ needs.verify.outputs.tag_name }}
+            type=sha,prefix=sha-
+
+      - name: Login to GHCR
+        uses: docker/login-action@v3
+        with:
+          registry: ghcr.io
+          username: ${{ github.actor }}
+          password: ${{ secrets.GITHUB_TOKEN }}
+
+      - name: Create manifest list and push
+        working-directory: /tmp/digests
+        run: |
+          docker buildx imagetools create \
+            $(jq -cr '.tags | map("-t " + .) | join(" ")' <<< "$DOCKER_METADATA_OUTPUT_JSON") \
+            $(printf 'ghcr.io/${{ github.repository_owner }}/multica-backend@sha256:%s ' *)
+
+      - name: Inspect image
+        run: |
+          docker buildx imagetools inspect \
+            ghcr.io/${{ github.repository_owner }}/multica-backend:${{ steps.meta.outputs.version }}
+
+  docker-web-build:
+    needs: verify
+    strategy:
+      fail-fast: false
+      matrix:
+        include:
+          - platform: linux/amd64
+            runs-on: ubuntu-latest
+          - platform: linux/arm64
+            runs-on: ubuntu-24.04-arm
+    runs-on: ${{ matrix.runs-on }}
+    steps:
+      - name: Prepare
+        run: |
+          platform=${{ matrix.platform }}
+          echo "PLATFORM_PAIR=${platform//\//-}" >> "$GITHUB_ENV"
+
+      - name: Checkout
+        uses: actions/checkout@v4
+
+      - name: Compute web image labels
+        id: meta
+        uses: docker/metadata-action@v5
+        with:
+          images: ghcr.io/${{ github.repository_owner }}/multica-web
+          labels: |
+            org.opencontainers.image.title=Multica Web
+            org.opencontainers.image.description=Multica self-hosted web frontend
+
+      - name: Setup Docker Buildx
+        uses: docker/setup-buildx-action@v3
+
+      - name: Login to GHCR
+        uses: docker/login-action@v3
+        with:
+          registry: ghcr.io
+          username: ${{ github.actor }}
+          password: ${{ secrets.GITHUB_TOKEN }}
+
+      - name: Build and push by digest
+        id: build
+        uses: docker/build-push-action@v6
+        with:
+          context: .
+          file: Dockerfile.web
+          pull: true
+          platforms: ${{ matrix.platform }}
+          labels: ${{ steps.meta.outputs.labels }}
+          cache-from: type=gha,scope=release-web-${{ env.PLATFORM_PAIR }}
+          cache-to: type=gha,mode=max,scope=release-web-${{ env.PLATFORM_PAIR }}
+          build-args: |
+            REMOTE_API_URL=http://backend:8080
+            NEXT_PUBLIC_APP_VERSION=${{ needs.verify.outputs.tag_name }}
+          outputs: type=image,name=ghcr.io/${{ github.repository_owner }}/multica-web,push-by-digest=true,name-canonical=true,push=true
+
+      - name: Export digest
+        run: |
+          mkdir -p /tmp/digests
+          digest="${{ steps.build.outputs.digest }}"
+          touch "/tmp/digests/${digest#sha256:}"
+
+      - name: Upload digest
+        uses: actions/upload-artifact@v4
+        with:
+          name: digests-web-${{ env.PLATFORM_PAIR }}
+          path: /tmp/digests/*
+          if-no-files-found: error
+          retention-days: 1
+
+  docker-web-merge:
+    needs: [verify, docker-web-build]
+    runs-on: ubuntu-latest
+    concurrency:
+      group: release-docker-web-${{ github.ref }}
+      cancel-in-progress: true
+    steps:
+      - name: Download digests
+        uses: actions/download-artifact@v4
+        with:
+          path: /tmp/digests
+          pattern: digests-web-*
+          merge-multiple: true
+
+      - name: Setup Docker Buildx
+        uses: docker/setup-buildx-action@v3
+
+      - name: Compute web image tags
+        id: meta
+        uses: docker/metadata-action@v5
+        with:
+          images: ghcr.io/${{ github.repository_owner }}/multica-web
+          flavor: |
+            latest=false
+          tags: |
+            type=raw,value=latest,enable=${{ needs.verify.outputs.is_stable == 'true' }}
+            type=raw,value=${{ needs.verify.outputs.tag_name }}
+            type=sha,prefix=sha-
+
+      - name: Login to GHCR
+        uses: docker/login-action@v3
+        with:
+          registry: ghcr.io
+          username: ${{ github.actor }}
+          password: ${{ secrets.GITHUB_TOKEN }}
+
+      - name: Create manifest list and push
+        working-directory: /tmp/digests
+        run: |
+          docker buildx imagetools create \
+            $(jq -cr '.tags | map("-t " + .) | join(" ")' <<< "$DOCKER_METADATA_OUTPUT_JSON") \
+            $(printf 'ghcr.io/${{ github.repository_owner }}/multica-web@sha256:%s ' *)
+
+      - name: Inspect image
+        run: |
+          docker buildx imagetools inspect \
+            ghcr.io/${{ github.repository_owner }}/multica-web:${{ steps.meta.outputs.version }}
+
+  # Build the Desktop installers for Linux and Windows and upload them to
+  # the GitHub Release that the `release` job above just published. macOS
+  # Desktop continues to ship via the manual `release-desktop` skill so it
+  # can be signed + notarized with Apple Developer credentials that are
+  # not (yet) wired into CI.
+  desktop:
+    needs: release
+    strategy:
+      fail-fast: false
+      matrix:
+        include:
+          - os: ubuntu-latest
+            target: linux
+          - os: windows-latest
+            target: win
+    runs-on: ${{ matrix.os }}
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v4
+        with:
+          fetch-depth: 0
+
+      - name: Install rpmbuild (Linux)
+        if: matrix.target == 'linux'
+        run: sudo apt-get update && sudo apt-get install -y rpm
+
+      - name: Setup Go
+        uses: actions/setup-go@v5
+        with:
+          go-version-file: server/go.mod
+          cache-dependency-path: server/go.sum
+
+      - name: Setup pnpm
+        uses: pnpm/action-setup@v4
+
+      - name: Setup Node
+        uses: actions/setup-node@v4
+        with:
+          node-version: 22
+          cache: pnpm
+
+      - name: Install dependencies
+        run: pnpm install --frozen-lockfile
+
+      - name: Package Desktop installers (${{ matrix.target }})
+        working-directory: apps/desktop
+        env:
+          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+          # electron-builder's GitHub publisher reads this:
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+          # Disable code signing on Linux/Windows for now — the public
+          # release is unsigned for these platforms, the CLI carries the
+          # trust boundary. Set CSC_LINK in repo secrets to enable
+          # Windows signing later.
+          CSC_IDENTITY_AUTO_DISCOVERY: "false"
+        run: node scripts/package.mjs --${{ matrix.target }} --x64 --arm64 --publish always

.gitignore

@@ -12,10 +12,17 @@ build
 bin
 dist-electron
 *.tsbuildinfo
+# ...except electron-builder's source resources dir, which holds tracked
+# config files (entitlements, icons) — not build output.
+!apps/desktop/build/
+!apps/desktop/build/**
 
 # env
 .env*
 !.env.example
+# Desktop production config is public (backend URL, etc.) — track it so
+# `pnpm package` produces a release-ready build without extra setup.
+!apps/desktop/.env.production
 
 # test coverage
 coverage
@@ -41,7 +48,13 @@ apps/web/test-results/
 # feature tracking
 _features/
 
+# runtime
+*.pid
+
 # platform specific
 *.dmg
 *.app
 server/server
+data/
+.kilo
+.idea

.goreleaser.yml

@@ -11,20 +11,39 @@ builds:
       - -s -w
       - -X main.version={{.Version}}
       - -X main.commit={{.ShortCommit}}
+      - -X main.date={{.Date}}
     env:
       - CGO_ENABLED=0
     goos:
       - darwin
       - linux
+      - windows
     goarch:
       - amd64
       - arm64
 
 archives:
-  - id: default
+  # Legacy archive name kept so already-released CLIs (whose `multica update`
+  # looks for `multica_{os}_{arch}.{ext}`) can keep self-updating. Remove
+  # once those versions are no longer in use.
+  - id: legacy
     formats:
       - tar.gz
+    format_overrides:
+      - goos: windows
+        formats:
+          - zip
     name_template: "{{ .ProjectName }}_{{ .Os }}_{{ .Arch }}"
+  # Versioned archive name used by current CLI / install scripts /
+  # desktop bootstrap going forward.
+  - id: versioned
+    formats:
+      - tar.gz
+    format_overrides:
+      - goos: windows
+        formats:
+          - zip
+    name_template: "{{ .ProjectName }}-cli-{{ .Version }}-{{ .Os }}-{{ .Arch }}"
 
 checksum:
   name_template: "checksums.txt"
@@ -39,6 +58,8 @@ changelog:
 
 brews:
   - name: multica
+    ids:
+      - versioned
     repository:
       owner: multica-ai
       name: homebrew-tap

.vercelignore

@@ -0,0 +1,85 @@
+# Deploy the frontend apps from the monorepo root.
+# Keep apps/web, apps/docs, shared packages, and root workspace metadata.
+# Exclude unrelated workspaces and local artifacts that can make
+# `vercel deploy` upload far more than the app needs.
+
+.agent_context
+.claude
+.context
+.env*
+.envrc
+.tool-versions
+_features
+.kilo
+.idea
+.DS_Store
+.husky
+.vscode
+
+/.dockerignore
+/.goreleaser.yml
+/AGENTS.md
+/CLAUDE.md
+/CLI_AND_DAEMON.md
+/CLI_INSTALL.md
+/CONTRIBUTING.md
+/Dockerfile
+/Dockerfile.web
+/HANDOFF_ARCHITECTURE_AUDIT.md
+/Makefile
+/README.md
+/README.zh-CN.md
+/SELF_HOSTING.md
+/SELF_HOSTING_ADVANCED.md
+/SELF_HOSTING_AI.md
+/docker-compose*.yml
+/playwright.config.ts
+/skills-lock.json
+
+/.github/
+/docker/
+/docs/
+/e2e/
+/server/
+/apps/desktop/
+/scripts/
+
+*.log
+*.pid
+*.tsbuildinfo
+
+.cache
+.next
+.pnpm-store
+.turbo
+.vercel
+coverage
+test-results
+playwright-report
+data
+
+node_modules
+bin
+dist
+out
+build
+dist-electron
+
+# Deployment-only trims: tests and lint configs are not used by `next build`.
+**/__tests__/**
+**/test/**
+**/*.test.*
+**/*.spec.*
+/packages/eslint-config/
+/apps/web/components.json
+/apps/web/eslint.config.mjs
+/apps/web/vitest.config.ts
+
+# Root repo metadata not needed in the deployment source.
+/.env.example
+/.gitattributes
+/.gitignore
+/LICENSE
+
+*.app
+*.dmg

AGENTS.md

@@ -2,273 +2,46 @@
 
 This file provides guidance to AI agents when working with code in this repository.
 
-## Project Context
+> **Single source of truth:** This file is a concise pointer document.
+> All authoritative architecture, coding rules, commands, and conventions
+> live in **CLAUDE.md** at the project root. Read that file first.
 
-Multica is an AI-native task management platform — like Linear, but with AI agents as first-class citizens.
+## Quick Reference
 
-- Agents can be assigned issues, create issues, comment, and change status
-- Supports local (daemon) and cloud agent runtimes
-- Built for 2-10 person AI-native teams
+### Architecture
 
-## Architecture
+Go backend + monorepo frontend (pnpm workspaces + Turborepo) with shared packages.
 
-**Go backend + standalone Next.js frontend.**
+- `server/` — Go backend (Chi router, sqlc, gorilla/websocket)
+- `apps/web/` — Next.js frontend (App Router)
+- `apps/desktop/` — Electron desktop app
+- `packages/core/` — Headless business logic (Zustand stores, React Query hooks, API client)
+- `packages/ui/` — Atomic UI components (shadcn/Base UI, zero business logic)
+- `packages/views/` — Shared business pages/components
+- `packages/tsconfig/` — Shared TypeScript config
 
-- `server/` — Go backend (Chi router, sqlc for DB, gorilla/websocket for real-time)
-- `apps/web/` — Next.js 16 frontend (App Router) — self-contained, no shared package dependencies
-- `e2e/` — Playwright end-to-end tests
-- `scripts/` and root `Makefile` — local setup and verification
+### State Management (critical)
 
-### Web App Structure (`apps/web/`)
+- **React Query** owns all server state (issues, members, agents, inbox, workspace list)
+- **Zustand** owns all client state (current workspace selection, view filters, drafts, modals)
+- All Zustand stores live in `packages/core/` — never in `packages/views/` or app directories
+- WS events invalidate React Query — never write directly to stores
 
-The frontend uses a **feature-based architecture** with four layers:
+### Package Boundaries (hard rules)
 
-```
-apps/web/
-├── app/          # Routing layer (thin shells — import from features/)
-├── features/     # Business logic, organized by domain
-├── shared/       # Cross-feature utilities (api client, types, logger)
-├── test/         # Shared test utilities and setup
-├── public/       # Static assets
-```
+- `packages/core/` — zero react-dom, zero localStorage, zero process.env
+- `packages/ui/` — zero `@multica/core` imports
+- `packages/views/` — zero `next/*`, zero `react-router-dom`, use `NavigationAdapter` for routing
+- `apps/web/platform/` — only place for Next.js APIs
 
-**`app/`** — Next.js App Router pages. Route files should be thin: import and re-export from `features/`. Layout components and route-specific glue (redirects, auth guards) live here. Shared layout components (e.g. `app-sidebar`) stay in `app/(dashboard)/_components/`.
-
-**`features/`** — Domain modules, each with its own components, hooks, stores, and config:
-
-| Feature | Purpose | Exports |
-|---|---|---|
-| `features/auth/` | Authentication state | `useAuthStore`, `AuthInitializer` |
-| `features/workspace/` | Workspace, members, agents | `useWorkspaceStore`, `useActorName` |
-| `features/issues/` | Issue state, components, config | `useIssueStore`, icons, pickers, status/priority config |
-| `features/inbox/` | Inbox notification state | `useInboxStore` |
-| `features/realtime/` | WebSocket connection + sync | `WSProvider`, `useWSEvent`, `useRealtimeSync` |
-| `features/modals/` | Modal registry and state | Modal store and components |
-| `features/skills/` | Skill management | Skill components |
-
-**`shared/`** — Code used across multiple features:
-- `shared/api/` — `ApiClient` (REST) and `WSClient` (WebSocket) for backend communication, plus the `api` singleton.
-- `shared/types/` — Domain types (Issue, Agent, Workspace, etc.) and WebSocket event types.
-- `shared/logger.ts` — Logger utility.
-
-### State Management
-
-- **Zustand** for global client state — one store per feature domain (`features/auth/store.ts`, `features/workspace/store.ts`, `features/issues/store.ts`, `features/inbox/store.ts`).
-- **React Context** only for connection lifecycle (`WSProvider` in `features/realtime/`).
-- **Local `useState`** for component-scoped UI state (forms, modals, filters).
-- Do not use React Context for data that can be a zustand store.
-
-**Store conventions:**
-- One store per feature domain. Import via `useAuthStore(selector)` or `useWorkspaceStore(selector)`.
-- Stores must not call `useRouter` or any React hooks — keep navigation in components.
-- Cross-store reads use `useOtherStore.getState()` inside actions (not hooks).
-- Dependency direction: `workspace` → `auth`, `realtime` → `auth`, `issues` → `workspace`. Never reverse.
-
-### Import Aliases
-
-Use `@/` alias (maps to `apps/web/`):
-```typescript
-import { api } from "@/shared/api";
-import type { Issue } from "@/shared/types";
-import { useAuthStore } from "@/features/auth";
-import { useWorkspaceStore } from "@/features/workspace";
-import { useIssueStore } from "@/features/issues";
-import { useInboxStore } from "@/features/inbox";
-import { useWSEvent } from "@/features/realtime";
-import { StatusIcon } from "@/features/issues/components";
-```
-
-Within a feature, use relative imports. Between features or to shared, use `@/`.
-
-### Data Flow
-
-```
-Browser → ApiClient (shared/api) → REST API (Chi handlers) → sqlc queries → PostgreSQL
-Browser ← WSClient (shared/api) ← WebSocket ← Hub.Broadcast() ← Handlers/TaskService
-```
-
-### Backend Structure (`server/`)
-
-- **Entry points** (`cmd/`): `server` (HTTP API), `multica` (CLI — daemon, agent management, config), `migrate`
-- **Handlers** (`internal/handler/`): One file per domain (issue, comment, agent, auth, daemon, etc.). Each handler holds `Queries`, `DB`, `Hub`, and `TaskService`.
-- **Real-time** (`internal/realtime/`): Hub manages WebSocket clients. Server broadcasts events; inbound WS message routing is still TODO.
-- **Auth** (`internal/auth/` + `internal/middleware/`): JWT (HS256). Middleware sets `X-User-ID` and `X-User-Email` headers. Login creates user on-the-fly if not found.
-- **Task lifecycle** (`internal/service/task.go`): Orchestrates agent work — enqueue → claim → start → complete/fail. Syncs issue status automatically and broadcasts WS events at each transition.
-- **Agent SDK** (`pkg/agent/`): Unified `Backend` interface for executing prompts via Claude Code or Codex. Each backend spawns its CLI and streams results via `Session.Messages` + `Session.Result` channels.
-- **Daemon** (`internal/daemon/`): Local agent runtime — auto-detects available CLIs (claude, codex), registers runtimes, polls for tasks, routes by provider.
-- **CLI** (`internal/cli/`): Shared helpers for the `multica` CLI — API client, config management, output formatting.
-- **Events** (`internal/events/`): Internal event bus for decoupled communication between handlers and services.
-- **Logging** (`internal/logger/`): Structured logging via slog. `LOG_LEVEL` env var controls level (debug, info, warn, error).
-- **Database**: PostgreSQL with pgvector extension (`pgvector/pgvector:pg17`). sqlc generates Go code from SQL in `pkg/db/queries/` → `pkg/db/generated/`. Migrations in `migrations/`.
-- **Routes** (`cmd/server/router.go`): Public routes (auth, health, ws) + protected routes (require JWT) + daemon routes (unauthenticated, separate auth model).
-
-### Multi-tenancy
-
-All queries filter by `workspace_id`. Membership checks gate access. `X-Workspace-ID` header routes requests to the correct workspace.
-
-### Agent Assignees
-
-Assignees are polymorphic — can be a member or an agent. `assignee_type` + `assignee_id` on issues. Agents render with distinct styling (purple background, robot icon).
-
-## Commands
+### Commands
 
 ```bash
-# One-click setup & run
-make setup            # First-time: ensure shared DB, create app DB, migrate
-make start            # Start backend + frontend together
-make stop             # Stop app processes for the current checkout
-make db-down          # Stop the shared PostgreSQL container
-
-# Frontend
-pnpm install
-pnpm dev:web          # Next.js dev server (port 3000)
-pnpm build            # Build frontend
+make dev              # Auto-setup + start everything
 pnpm typecheck        # TypeScript check
-pnpm lint             # ESLint via Next.js
-pnpm test             # TS tests (Vitest)
-
-# Backend (Go)
-make dev              # Run Go server (port 8080)
-make daemon           # Run local daemon
-make build            # Build server + CLI binaries to server/bin/
-make cli ARGS="..."   # Run multica CLI (e.g. make cli ARGS="config")
+pnpm test             # TS unit tests (Vitest)
 make test             # Go tests
-make sqlc             # Regenerate sqlc code after editing SQL in server/pkg/db/queries/
-make migrate-up       # Run database migrations
-make migrate-down     # Rollback migrations
-
-# Run a single Go test
-cd server && go test ./internal/handler/ -run TestName
-
-# Run a single TS test
-pnpm --filter @multica/web exec vitest run src/path/to/file.test.ts
-
-# Run a single E2E test (requires backend + frontend running)
-pnpm exec playwright test e2e/tests/specific-test.spec.ts
-
-# Infrastructure
-make db-up            # Start shared PostgreSQL (pgvector/pg17 image)
-make db-down          # Stop shared PostgreSQL
+make check            # Full verification pipeline
 ```
 
-### CI Requirements
-
-CI runs on Node 22 and Go 1.24 with a `pgvector/pgvector:pg17` PostgreSQL service. See `.github/workflows/ci.yml`.
-
-### Worktree Support
-
-All checkouts share one PostgreSQL container. Isolation is at the database level — each worktree gets its own DB name and unique ports via `.env.worktree`. Main checkouts use `.env`.
-
-```bash
-make worktree-env       # Generate .env.worktree with unique DB/ports
-make setup-worktree     # Setup using .env.worktree
-make start-worktree     # Start using .env.worktree
-```
-
-## Coding Rules
-
-- TypeScript strict mode is enabled; keep types explicit.
-- TypeScript in `apps/web` uses 2-space indentation, double quotes, and semicolons.
-- Prefer PascalCase for React components, camelCase for hooks and helpers, and colocated test files such as `page.test.tsx`.
-- Go code follows standard Go conventions (gofmt, go vet). Use domain-oriented filenames like `issue.go` or `cmd_issue.go`.
-- Do not hand-edit generated code in `server/pkg/db/generated/`.
-- Keep comments in code **English only**.
-- Prefer existing patterns/components over introducing parallel abstractions.
-- Unless the user explicitly asks for backwards compatibility, do **not** add compatibility layers, fallback paths, dual-write logic, legacy adapters, or temporary shims.
-- If a flow or API is being replaced and the product is not yet live, prefer removing the old path instead of preserving both old and new behavior.
-- Treat compatibility code as a maintenance cost, not a default safety mechanism. Avoid "just in case" branches that make the codebase harder to reason about.
-- Avoid broad refactors unless required by the task.
-
-## UI/UX Rules
-
-- Prefer shadcn components over custom implementations. Install missing components via `npx shadcn add`.
-- **Feature-specific components** → `features/<domain>/components/` — issue icons, pickers, and other domain-bound UI live inside their feature module.
-- Use shadcn design tokens for styling (e.g. `bg-primary`, `text-muted-foreground`, `text-destructive`). Avoid hardcoded color values (e.g. `text-red-500`, `bg-gray-100`).
-- Do not introduce extra state (useState, context, reducers) unless explicitly required by the design. Prefer zustand stores for shared state over React Context.
-- Pay close attention to **overflow** (truncate long text, scrollable containers), **alignment**, and **spacing** consistency.
-- When unsure about interaction or state design, ask — the user will provide direction.
-
-## Testing Rules
-
-- **TypeScript**: Vitest with Testing Library. Shared test setup lives in `apps/web/test/`. Mock external/third-party dependencies only.
-- **Go**: Standard `go test`. Tests should create their own fixture data in a test database.
-- End-to-end tests live in `e2e/*.spec.ts`; `make check` will start missing services automatically, while direct Playwright runs expect the app to already be running.
-- Add or update tests whenever you change handlers, CLI commands, daemon behavior, or SQL-backed flows.
-
-## Commit & Pull Request Rules
-
-- Use atomic commits grouped by logical intent.
-- Conventional format with scopes:
-  - `feat(web): ...`, `feat(cli): ...`
-  - `fix(web): ...`, `fix(cli): ...`
-  - `refactor(daemon): ...`
-  - `test(cli): ...`
-  - `docs: ...`
-  - `chore(scope): ...`
-- Keep PRs focused and include a short description, linked issue or PR number when relevant, screenshots for UI work, and notes for migrations, env changes, or CLI surface changes.
-- Before opening a PR, run `make check` or the relevant frontend/backend subset.
-
-## Minimum Pre-Push Checks
-
-```bash
-make check    # Runs all checks: typecheck, unit tests, Go tests, E2E
-```
-
-Run verification only when the user explicitly asks for it.
-
-For targeted checks when requested:
-```bash
-pnpm typecheck        # TypeScript type errors only
-pnpm test             # TS unit tests only (Vitest)
-make test             # Go tests only
-pnpm exec playwright test   # E2E only (requires backend + frontend running)
-```
-
-## AI Agent Verification Loop
-
-After writing or modifying code, always run the full verification pipeline:
-
-```bash
-make check
-```
-
-This runs all checks in sequence:
-1. TypeScript typecheck (`pnpm typecheck`)
-2. TypeScript unit tests (`pnpm test`)
-3. Go tests (`go test ./...`)
-4. E2E tests (auto-starts backend + frontend if needed, runs Playwright)
-
-**Workflow:**
-- Write code to satisfy the requirement
-- Run `make check`
-- If any step fails, read the error output, fix the code, and re-run `make check`
-- Repeat until all checks pass
-- Only then consider the task complete
-
-**Quick iteration:** If you know only TypeScript or Go is affected, run individual checks first for faster feedback, then finish with a full `make check` before marking work complete.
-
-## E2E Test Patterns
-
-E2E tests should be self-contained. Use the `TestApiClient` fixture for data setup/teardown:
-
-```typescript
-import { loginAsDefault, createTestApi } from "./helpers";
-import type { TestApiClient } from "./fixtures";
-
-let api: TestApiClient;
-
-test.beforeEach(async ({ page }) => {
-  api = await createTestApi();       // logged-in API client
-  await loginAsDefault(page);        // browser session
-});
-
-test.afterEach(async () => {
-  await api.cleanup();               // delete any data created during the test
-});
-
-test("example", async ({ page }) => {
-  const issue = await api.createIssue("Test Issue");  // create via API
-  await page.goto(`/issues/${issue.id}`);             // test via UI
-  // api.cleanup() in afterEach removes the issue
-});
-```
+See CLAUDE.md for the complete command reference.

CLAUDE.md

@@ -2,6 +2,21 @@
 
 This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.
 
+## Conventions reference
+
+The single source of truth for **code naming, the i18n translation glossary, and the Chinese voice guide** is the docs site:
+
+- **`apps/docs/content/docs/developers/conventions.mdx`** (English)
+- **`apps/docs/content/docs/developers/conventions.zh.mdx`** (Chinese)
+
+Read that page before:
+
+- Writing or editing translations (`packages/views/locales/`)
+- Naming a new route, package, file, DB column, or TS type
+- Writing Chinese product copy (UI strings, error messages, docs)
+
+The legacy `packages/views/locales/glossary.md` is now a stub redirecting to the docs page; do not rely on it.
+
 ## Project Context
 
 Multica is an AI-native task management platform — like Linear, but with AI agents as first-class citizens.
@@ -12,119 +27,71 @@ Multica is an AI-native task management platform — like Linear, but with AI ag
 
 ## Architecture
 
-**Go backend + standalone Next.js frontend.**
+**Go backend + monorepo frontend (pnpm workspaces + Turborepo) with shared packages.**
 
 - `server/` — Go backend (Chi router, sqlc for DB, gorilla/websocket for real-time)
-- `apps/web/` — Next.js 16 frontend (App Router) — self-contained, no shared package dependencies
+- `apps/web/` — Next.js frontend (App Router)
+- `apps/desktop/` — Electron desktop app (electron-vite)
+- `packages/core/` — Headless business logic (zero react-dom, all-platform reuse)
+- `packages/ui/` — Atomic UI components (zero business logic)
+- `packages/views/` — Shared business pages/components (zero next/* imports, zero react-router imports)
+- `packages/tsconfig/` — Shared TypeScript configuration
 
-### Web App Structure (`apps/web/`)
+### Key Architectural Decisions
 
-The frontend uses a **feature-based architecture** with four layers:
+**Internal Packages pattern** — all shared packages export raw `.ts`/`.tsx` files (no pre-compilation). The consuming app's bundler compiles them directly. This gives zero-config HMR and instant go-to-definition.
 
-```
-apps/web/
-├── app/          # Routing layer (thin shells — import from features/)
-├── features/     # Business logic, organized by domain
-├── shared/       # Cross-feature utilities (api client, types, logger)
-```
+**Dependency direction:** `views/ → core/ + ui/`. Core and UI are independent of each other. No package imports from `next/*`, `react-router-dom`, or app-specific code.
 
-**`app/`** — Next.js App Router pages. Route files should be thin: import and re-export from `features/`. Layout components and route-specific glue (redirects, auth guards) live here. Shared layout components (e.g. `app-sidebar`) stay in `app/(dashboard)/_components/`.
+**Platform bridge:** `packages/core/platform/` provides `CoreProvider` — initializes API client, auth/workspace stores, WS connection, and QueryClient. Each app wraps its root with `<CoreProvider>` and provides its own `NavigationAdapter` for routing.
 
-**`features/`** — Domain modules, each with its own components, hooks, stores, and config:
-
-| Feature | Purpose | Exports |
-|---|---|---|
-| `features/auth/` | Authentication state | `useAuthStore`, `AuthInitializer` |
-| `features/workspace/` | Workspace, members, agents | `useWorkspaceStore`, `useActorName` |
-| `features/issues/` | Issue state, components, config | `useIssueStore`, icons, pickers, status/priority config |
-| `features/inbox/` | Inbox notification state | `useInboxStore` |
-| `features/realtime/` | WebSocket connection + sync | `WSProvider`, `useWSEvent`, `useRealtimeSync` |
-| `features/modals/` | Modal registry and state | Modal store and components |
-| `features/skills/` | Skill management | Skill components |
-
-**`shared/`** — Code used across multiple features:
-- `shared/api/` — `ApiClient` (REST) and `WSClient` (WebSocket) for backend communication, plus the `api` singleton.
-- `shared/types/` — Domain types (Issue, Agent, Workspace, etc.) and WebSocket event types.
-- `shared/logger.ts` — Logger utility.
+**pnpm catalog** — `pnpm-workspace.yaml` defines `catalog:` for version pinning. All shared deps use `catalog:` references to guarantee a single version across all packages. When adding new shared deps (including test deps), add to catalog first.
 
 ### State Management
 
-- **Zustand** for global client state — one store per feature domain (`features/auth/store.ts`, `features/workspace/store.ts`, `features/issues/store.ts`, `features/inbox/store.ts`).
-- **React Context** only for connection lifecycle (`WSProvider` in `features/realtime/`).
-- **Local `useState`** for component-scoped UI state (forms, modals, filters).
-- Do not use React Context for data that can be a zustand store.
+The architecture relies on a strict split between server state and client state. Mixing them is the most common way to break it.
 
-**Store conventions:**
-- One store per feature domain. Import via `useAuthStore(selector)` or `useWorkspaceStore(selector)`.
-- Stores must not call `useRouter` or any React hooks — keep navigation in components.
-- Cross-store reads use `useOtherStore.getState()` inside actions (not hooks).
-- Dependency direction: `workspace` → `auth`, `realtime` → `auth`, `issues` → `workspace`. Never reverse.
+- **TanStack Query owns all server state.** Issues, users, workspaces, inbox — anything fetched from the API lives in the Query cache. WS events keep it fresh via invalidation; no polling, no `staleTime` workarounds.
+- **Zustand owns all client state.** UI selections, filters, drafts, modal state, navigation history. Stores live in `packages/core/` (never in `packages/views/`) so both apps share them.
+- **React Context** is reserved for cross-cutting platform plumbing — `WorkspaceIdProvider`, `NavigationProvider`. Don't reach for it for general state.
+- **Auth and workspace stores are the only stores allowed to call `api.*` directly**, because they manage critical state that must exist before queries can run. They're created via factory + injected dependencies, registered by the platform layer.
 
-### Import Aliases
+**Hard rules — these are how the architecture stays coherent:**
 
-Use `@/` alias (maps to `apps/web/`):
-```typescript
-import { api } from "@/shared/api";
-import type { Issue } from "@/shared/types";
-import { useAuthStore } from "@/features/auth";
-import { useWorkspaceStore } from "@/features/workspace";
-import { useIssueStore } from "@/features/issues";
-import { useInboxStore } from "@/features/inbox";
-import { useWSEvent } from "@/features/realtime";
-import { StatusIcon } from "@/features/issues/components";
-```
+- **Never duplicate server data into Zustand.** If it came from the API, it belongs in the Query cache. Copying it into a store creates two sources of truth and they will drift.
+- **Workspace-scoped queries must key on `wsId`.** This is what makes workspace switching automatic — the cache key changes, the right data appears, no manual invalidation needed.
+- **Mutations are optimistic by default.** Apply the change locally, send the request, roll back on failure, invalidate on settle. The user shouldn't wait for the server.
+- **WS events invalidate queries — they never write to stores directly.** This keeps the cache as the single source of truth and avoids race conditions.
+- **Persist what's worth preserving across restarts** (user preferences, drafts, tab layout). **Don't persist ephemeral UI state** (modal open/close, transient selections) or server data.
 
-Within a feature, use relative imports. Between features or to shared, use `@/`.
+**Common Zustand footguns to avoid:**
 
-### Data Flow
-
-```
-Browser → ApiClient (shared/api) → REST API (Chi handlers) → sqlc queries → PostgreSQL
-Browser ← WSClient (shared/api) ← WebSocket ← Hub.Broadcast() ← Handlers/TaskService
-```
-
-### Backend Structure (`server/`)
-
-- **Entry points** (`cmd/`): `server` (HTTP API), `multica` (CLI — daemon, agent management, config), `migrate`
-- **Handlers** (`internal/handler/`): One file per domain (issue, comment, agent, auth, daemon, etc.). Each handler holds `Queries`, `DB`, `Hub`, and `TaskService`.
-- **Real-time** (`internal/realtime/`): Hub manages WebSocket clients. Server broadcasts events; inbound WS message routing is still TODO.
-- **Auth** (`internal/auth/` + `internal/middleware/`): JWT (HS256). Middleware sets `X-User-ID` and `X-User-Email` headers. Login creates user on-the-fly if not found.
-- **Task lifecycle** (`internal/service/task.go`): Orchestrates agent work — enqueue → claim → start → complete/fail. Syncs issue status automatically and broadcasts WS events at each transition.
-- **Agent SDK** (`pkg/agent/`): Unified `Backend` interface for executing prompts via Claude Code or Codex. Each backend spawns its CLI and streams results via `Session.Messages` + `Session.Result` channels.
-- **Daemon** (`internal/daemon/`): Local agent runtime — auto-detects available CLIs (claude, codex), registers runtimes, polls for tasks, routes by provider.
-- **CLI** (`internal/cli/`): Shared helpers for the `multica` CLI — API client, config management, output formatting.
-- **Events** (`internal/events/`): Internal event bus for decoupled communication between handlers and services.
-- **Logging** (`internal/logger/`): Structured logging via slog. `LOG_LEVEL` env var controls level (debug, info, warn, error).
-- **Database**: PostgreSQL with pgvector extension (`pgvector/pgvector:pg17`). sqlc generates Go code from SQL in `pkg/db/queries/` → `pkg/db/generated/`. Migrations in `migrations/`.
-- **Routes** (`cmd/server/router.go`): Public routes (auth, health, ws) + protected routes (require JWT) + daemon routes (unauthenticated, separate auth model).
-
-### Multi-tenancy
-
-All queries filter by `workspace_id`. Membership checks gate access. `X-Workspace-ID` header routes requests to the correct workspace.
-
-### Agent Assignees
-
-Assignees are polymorphic — can be a member or an agent. `assignee_type` + `assignee_id` on issues. Agents render with distinct styling (purple background, robot icon).
+- Selectors must return stable references. Returning a freshly built object or array on every call (e.g. `s => ({ a: s.a, b: s.b })` or `s => s.items.map(...)`) triggers infinite re-renders. Either select primitives separately or use shallow comparison.
+- Hooks that need workspace context should accept `wsId` as a parameter, not call `useWorkspaceId()` internally — this lets them work outside the `WorkspaceIdProvider` (e.g. in a sidebar that renders before workspace is loaded).
 
 ## Commands
 
 ```bash
-# One-click setup & run
+# One-command dev (auto-setup + start everything)
+make dev              # Auto-creates env, installs deps, starts DB, migrates, launches app
+
+# Explicit setup & run (if you prefer separate steps)
 make setup            # First-time: ensure shared DB, create app DB, migrate
 make start            # Start backend + frontend together
 make stop             # Stop app processes for the current checkout
 make db-down          # Stop the shared PostgreSQL container
 
-# Frontend
+# Frontend (all commands go through Turborepo)
 pnpm install
 pnpm dev:web          # Next.js dev server (port 3000)
-pnpm build            # Build frontend
-pnpm typecheck        # TypeScript check
-pnpm lint             # ESLint via Next.js
-pnpm test             # TS tests (Vitest)
+pnpm dev:desktop      # Electron dev (electron-vite, HMR)
+pnpm build            # Build all frontend apps
+pnpm typecheck        # TypeScript check (all packages + apps via turbo)
+pnpm lint             # ESLint
+pnpm test             # TS tests (Vitest, all packages + apps via turbo)
 
 # Backend (Go)
-make dev              # Run Go server (port 8080)
+make server           # Run Go server only (port 8080)
 make daemon           # Run local daemon
 make build            # Build server + CLI binaries to server/bin/
 make cli ARGS="..."   # Run multica CLI (e.g. make cli ARGS="config")
@@ -133,18 +100,28 @@ make sqlc             # Regenerate sqlc code after editing SQL in server/pkg/db/
 make migrate-up       # Run database migrations
 make migrate-down     # Rollback migrations
 
+# Run a single TS test (works for any package with a test script)
+pnpm --filter @multica/views exec vitest run auth/login-page.test.tsx
+pnpm --filter @multica/core exec vitest run runtimes/version.test.ts
+pnpm --filter @multica/web exec vitest run app/\(auth\)/login/page.test.tsx
+
 # Run a single Go test
 cd server && go test ./internal/handler/ -run TestName
 
-# Run a single TS test
-pnpm --filter @multica/web exec vitest run src/path/to/file.test.ts
-
 # Run a single E2E test (requires backend + frontend running)
 pnpm exec playwright test e2e/tests/specific-test.spec.ts
 
+# Desktop build & package
+pnpm --filter @multica/desktop build      # Compile TS → JS (reads .env.production)
+pnpm --filter @multica/desktop package    # Package into .app/.dmg/.exe (current platform only)
+
+# shadcn — config lives in packages/ui/components.json (Base UI variant, base-nova style)
+pnpm ui:add badge                # Adds component to packages/ui/components/ui/
+
 # Infrastructure
 make db-up            # Start shared PostgreSQL (pgvector/pg17 image)
 make db-down          # Stop shared PostgreSQL
+make db-reset         # Drop + recreate current env's DB, then re-run migrations (local only; stop backend first)
 ```
 
 ### CI Requirements
@@ -155,6 +132,8 @@ CI runs on Node 22 and Go 1.26.1 with a `pgvector/pgvector:pg17` PostgreSQL serv
 
 All checkouts share one PostgreSQL container. Isolation is at the database level — each worktree gets its own DB name and unique ports via `.env.worktree`. Main checkouts use `.env`.
 
+`make dev` auto-detects worktrees and handles everything. For explicit control:
+
 ```bash
 make worktree-env       # Generate .env.worktree with unique DB/ports
 make setup-worktree     # Setup using .env.worktree
@@ -167,45 +146,197 @@ make start-worktree     # Start using .env.worktree
 - Go code follows standard Go conventions (gofmt, go vet).
 - Keep comments in code **English only**.
 - Prefer existing patterns/components over introducing parallel abstractions.
-- Unless the user explicitly asks for backwards compatibility, do **not** add compatibility layers, fallback paths, dual-write logic, legacy adapters, or temporary shims.
+- Unless the user explicitly asks for backwards compatibility, do **not** add compatibility layers, fallback paths, dual-write logic, legacy adapters, or temporary shims **for internal, non-boundary code** (a function calling another function in the same package, a component reading its own state, a store helper, etc.).
+- This rule does **not** apply at API boundaries: the desktop app cannot assume the backend it talks to has the same shape as the one it was built against (older desktop installs will outlive any given server build). API response handling must follow the rules in **API Response Compatibility** below — that is a defensive boundary, not a legacy shim.
 - If a flow or API is being replaced and the product is not yet live, prefer removing the old path instead of preserving both old and new behavior.
-- Treat compatibility code as a maintenance cost, not a default safety mechanism. Avoid "just in case" branches that make the codebase harder to reason about.
 - Avoid broad refactors unless required by the task.
+- New global (pre-workspace) routes MUST use a single word (`/login`, `/inbox`) or a `/{noun}/{verb}` pair (`/workspaces/new`). NEVER add hyphenated word-group root routes (`/new-workspace`, `/create-team`) — they collide with common user workspace names and force endless reserved-slug audits. Reserving the noun (`workspaces`) automatically protects the entire `/workspaces/*` subtree.
+- The reserved-slug list lives in **one** place: `server/internal/handler/reserved_slugs.json`. The Go side embeds the JSON; `packages/core/paths/reserved-slugs.ts` is generated from it by `pnpm generate:reserved-slugs`. Edit the JSON, run the generator, commit both. CI re-runs the generator and fails on any drift, so a stale TS file cannot land.
+
+### API Response Compatibility
+
+The desktop app installed on a user's machine is older than any backend it talks to: a user on 0.2.26 will hit a server running 0.3.x, then 0.4.x, then beyond. Every response shape is a contract that **will** drift, and the frontend must survive drift without white-screening. Three concrete incidents already happened from violating this — #2143, #2147, #2192.
+
+When writing code that consumes an API response, follow these rules:
+
+- **Parse, don't cast.** Untyped JSON crossing the network is not `T`. Use `parseWithFallback` in `packages/core/api/schema.ts` with a `zod` schema and an explicit fallback. On validation failure it logs a warning and returns the fallback; it never throws into the UI.
+- **No bare `as` casts on response bodies.** Every endpoint method whose response is consumed by UI logic must run through a schema before returning.
+- **Optional-chain and default everywhere downstream.** Treat every field as possibly missing. Use explicit boolean checks (`=== true`) over truthy/falsy negation, which silently treats `undefined` and `null` as `false`.
+- **Don't pin a UI affordance to a single backend field.** If a button or indicator depends on exactly one boolean from the server, a backend bug deletes it. Combine signals (cursor presence, page length, etc.) so the affordance stays available in the worst case.
+- **Enum drift downgrades, not crashes.** A new server-side enum value should render a generic fallback. `switch` statements on server-driven strings must have a `default` branch.
+- **When you add or change an endpoint:** add the schema in the same PR, and write at least one test that feeds a malformed response through it (missing field, wrong type, `null` array). The test fails closed if a future change breaks the contract.
+
+This is not premature defense — it is the *only* defense for an installed-app architecture. CSR-only browser apps can ship a fix in minutes; an Electron build sitting on a developer's laptop cannot.
+
+### Backend Handler UUID Parsing Convention
+
+Every Go handler in `server/internal/handler/` follows these rules. The convention exists because `util.ParseUUID` used to silently return a zero UUID on invalid input, which caused #1661 — a `DELETE` returning 204 success while the SQL `DELETE` matched zero rows.
+
+- **Resource path params that accept either a UUID or a human-readable identifier** (e.g. `chi.URLParam(r, "id")` for an issue, which accepts both `MUL-123` and a UUID) MUST be resolved through the dedicated loader (`loadIssueForUser` / `loadSkillForUser` / `loadAgentForUser` / `requireDaemonRuntimeAccess`). After resolution, all subsequent DB calls — especially `Queries.Delete*` / `Queries.Update*` — MUST use `entity.ID` from the resolved object. Never round-trip the raw URL string through `parseUUID` for a write query.
+- **Pure-UUID inputs from request boundaries** (URL params that are always UUIDs, request body fields, query params, headers) MUST be validated with `parseUUIDOrBadRequest(w, s, fieldName)`. On invalid input it writes a 400 and returns `ok=false` — return immediately.
+- **Trusted UUID round-trips** (sqlc-returned UUIDs being passed back into queries, test fixtures) use `parseUUID(s)` which calls `util.MustParseUUID` and panics on invalid input. A panic here means an unguarded user-input string slipped in — that is a real bug. `chi`'s `middleware.Recoverer` translates the panic into a 500 so the process keeps running.
+- **`util.ParseUUID(s) (pgtype.UUID, error)`** is the only safe variant outside the handler package. Always check the error.
+
+When adding a `Queries.Delete*` or `Queries.Update*` call, ask: "Where did this UUID come from?" If the answer is "raw user input that hasn't been validated," route it through `parseUUIDOrBadRequest` or a loader first.
+
+### Package Boundary Rules
+
+These are hard constraints. Violating them breaks the cross-platform architecture:
+
+- `packages/core/` — zero react-dom, zero localStorage (use StorageAdapter), zero process.env, zero UI libraries. **All shared Zustand stores live here**, even view-related ones (filters, view modes) — stores are pure state, not UI.
+- `packages/ui/` — zero `@multica/core` imports (pure UI, no business logic).
+- `packages/views/` — zero `next/*` imports, zero `react-router-dom` imports, zero stores. Use `NavigationAdapter` for all routing.
+- `apps/web/platform/` — the only place for Next.js APIs (`next/navigation`).
+- `apps/desktop/src/renderer/src/platform/` — the only place for react-router-dom navigation wiring.
+
+### The No-Duplication Rule
+
+**If the same logic exists in both apps, it must be extracted to a shared package.**
+
+This applies to everything: components, hooks, guards, providers, utility functions. The decision process:
+
+1. Does this code depend on Next.js or Electron APIs? → Keep in the respective app.
+2. Does it depend on `react-router-dom` or `next/navigation`? → Keep in app's `platform/` layer.
+3. Everything else → belongs in `packages/core/` (headless logic) or `packages/views/` (UI components).
+
+When the two apps need different behavior for the same concept (e.g., different loading UI), extract the shared logic into a component with props/slots for the differences. Don't duplicate the logic.
+
+### Cross-Platform Development Rules
+
+When adding a new page or feature:
+
+1. **New page component** → add to `packages/views/<domain>/`. Never import from `next/*` or `react-router-dom`.
+2. **Wire it in both apps** → add a route in `apps/web/app/` (Next.js page file) AND in the desktop router. **Exception**: pre-workspace transition flows (create workspace, accept invite) are NOT routes on desktop — they're `WindowOverlay` state. See *Desktop-specific Rules → Route categories*.
+3. **Navigation** → use `useNavigation().push()` or `<AppLink>`. Never use framework-specific link/router APIs in shared code.
+4. **Shared guards/providers** → use `DashboardGuard` from `packages/views/layout/`. Don't create separate guard logic per app.
+5. **Platform-specific UI** → if a feature is web-only or desktop-only, keep it in the respective app. Use props slots (`extra`, `topSlot`) on shared layout components to inject platform-specific UI.
+6. **New hooks that need workspace context** → accept `wsId` as parameter instead of reading from `useWorkspaceId()` Context, so they work both inside and outside `WorkspaceIdProvider`.
+
+### CSS Architecture
+
+Both apps share the same CSS foundation from `packages/ui/styles/`.
+
+- **Design tokens** → use semantic tokens (`bg-background`, `text-muted-foreground`). Never use hardcoded Tailwind colors (`text-red-500`, `bg-gray-100`).
+- **Shared styles** → `packages/ui/styles/`. Never duplicate scrollbar styling, keyframes, or base layer rules in app CSS.
+- **`@source` directives** → both apps scan shared packages so Tailwind sees all class names.
+
+## Desktop-specific Rules
+
+These rules apply to `apps/desktop/` only. Web has different constraints (URL bar, SSR, no tabs) and doesn't share these concerns. Every rule in this section was added after a concrete bug — treat them as enforced, not suggestions.
+
+### Route categories
+
+Every path in the desktop app falls into exactly one category. Choosing the wrong one reproduces bugs we've already fixed.
+
+- **Session routes** — workspace-scoped pages (`/:slug/issues`, `/:slug/settings`). Rendered by the per-tab memory router under `WorkspaceRouteLayout`. These are legitimate tab destinations.
+- **Transition flows** — pre-workspace / one-shot actions (create workspace, accept invite). **NOT routes.** They live as `WindowOverlay` state, dispatched when the navigation adapter sees `push('/workspaces/new')` or `push('/invite/<id>')`. The shared view (`NewWorkspacePage`, `InvitePage`) is the content; the overlay wrapper supplies platform chrome.
+- **Error / stale states** — "workspace not available", tabs pointing at a revoked workspace. **NOT pages.** `WorkspaceRouteLayout` auto-heals by dropping the stale tab group from the store; the user never lands on an explicit error screen. Web keeps `NoAccessPage` (shareable URL makes the error state meaningful); desktop has no URL bar so stale = heal silently.
+
+**Adding a new pre-workspace flow on desktop**: register a new `WindowOverlay` type in `stores/window-overlay-store.ts`. Do NOT add it to `routes.tsx`. If a shared view needs the flow on both platforms, add the route on web (`apps/web/app/(auth)/...`) AND the overlay type on desktop — the shared view component is identical.
+
+### Workspace context
+
+`setCurrentWorkspace(slug, uuid)` from `@multica/core/platform` is the single source of truth for the active workspace. `WorkspaceRouteLayout` sets it on mount; unmount does NOT clear it. Code that leaves workspace context (leave/delete workspace, force-navigate to overlay) must call `setCurrentWorkspace(null, null)` explicitly.
+
+### Workspace destructive operations
+
+Leave / Delete workspace flows must follow this order, otherwise concurrent refetches race and the renderer hard-reloads:
+
+1. Read destination from cached workspace list.
+2. `setCurrentWorkspace(null, null)`.
+3. `navigation.push(destination)`.
+4. THEN `await mutation.mutateAsync(workspaceId)`.
+
+### Tab isolation
+
+Tabs are grouped per workspace in `stores/tab-store.ts`. The TabBar shows only the active workspace's tabs; cross-workspace tab leakage is impossible by construction (no flat global tabs array).
+
+Cross-workspace `push(path)` is detected by the navigation adapter (`platform/navigation.tsx`) and translated into `switchWorkspace(slug, targetPath)` — NOT a navigation within the current tab's router. Don't bypass the adapter; always go through `useNavigation()` from shared code.
+
+### Drag region (macOS)
+
+Every full-window desktop view (anything outside the dashboard shell) must mount `<DragStrip />` from `@multica/views/platform` as the first flex child of the page root, otherwise users can't drag the window. Interactive UI inside the top 48px needs `WebkitAppRegion: "no-drag"` to stay clickable.
 
 ## UI/UX Rules
 
-- Prefer shadcn components over custom implementations. Install missing components via `npx shadcn add`.
-- **Feature-specific components** → `features/<domain>/components/` — issue icons, pickers, and other domain-bound UI live inside their feature module.
-- Use shadcn design tokens for styling (e.g. `bg-primary`, `text-muted-foreground`, `text-destructive`). Avoid hardcoded color values (e.g. `text-red-500`, `bg-gray-100`).
-- Do not introduce extra state (useState, context, reducers) unless explicitly required by the design. Prefer zustand stores for shared state over React Context.
+- Prefer shadcn components over custom implementations. Install via `pnpm ui:add <component>` from project root — adds to `packages/ui/components/ui/`. All components use Base UI primitives (`@base-ui/react`), not Radix.
+- Use shadcn design tokens for styling. Avoid hardcoded color values.
+- Do not introduce extra state (useState, context, reducers) unless explicitly required by the design.
 - Pay close attention to **overflow** (truncate long text, scrollable containers), **alignment**, and **spacing** consistency.
-- When unsure about interaction or state design, ask — the user will provide direction.
+- **If a component is identical between web and desktop, it belongs in a shared package.** Do not copy-paste between apps.
 
 ## Testing Rules
 
-- **TypeScript**: Vitest. Mock external/third-party dependencies only.
-- **Go**: Standard `go test`. Tests should create their own fixture data in a test database.
+### Where to write tests
+
+Tests follow the code, not the app. This is the most important testing principle in this monorepo:
+
+| What you're testing | Where the test lives | Why |
+|---|---|---|
+| Shared business logic (stores, queries, hooks) | `packages/core/*.test.ts` | No DOM needed, pure logic |
+| Shared UI components (pages, forms, modals) | `packages/views/*.test.tsx` | jsdom, no framework mocks |
+| Platform-specific wiring (cookies, redirects, searchParams) | `apps/web/*.test.tsx` or `apps/desktop/` | Needs framework-specific mocks |
+| End-to-end user flows | `e2e/*.spec.ts` | Real browser, real backend |
+
+**Never test shared component behavior in an app's test file.** If a test requires mocking `next/navigation` or `react-router-dom` to test a component from `@multica/views`, the test is in the wrong place — move it to `packages/views/` and mock `@multica/core` instead.
+
+### Test infrastructure
+
+- `packages/core/` — Vitest, Node environment (no DOM)
+- `packages/views/` — Vitest, jsdom environment, `@testing-library/react`
+- `apps/web/` — Vitest, jsdom environment, framework-specific mocks
+- `e2e/` — Playwright
+- `server/` — Go standard `go test`
+
+All test deps are in the pnpm catalog for unified versioning.
+
+### Mocking conventions
+
+- Mock `@multica/core` stores with `vi.hoisted()` + `Object.assign(selectorFn, { getState })` pattern (Zustand stores are both callable and have `.getState()`).
+- Mock `@multica/core/api` for API calls.
+- In `packages/views/` tests: never mock `next/*` or `react-router-dom` — those don't exist here.
+- In `apps/web/` tests: mock framework-specific APIs only for platform-specific behavior.
+
+### TDD workflow
+
+1. Write failing test in the **correct package** first.
+2. Write implementation.
+3. Run `pnpm test` (Turborepo discovers all packages).
+4. Green → done.
+
+### Go tests
+
+Standard `go test`. Tests should create their own fixture data in a test database.
+
+### E2E tests
+
+E2E tests should be self-contained. Use the `TestApiClient` fixture for data setup/teardown:
+
+```typescript
+import { loginAsDefault, createTestApi } from "./helpers";
+import type { TestApiClient } from "./fixtures";
+
+let api: TestApiClient;
+
+test.beforeEach(async ({ page }) => {
+  api = await createTestApi();
+  await loginAsDefault(page);
+});
+
+test.afterEach(async () => {
+  await api.cleanup();
+});
+
+test("example", async ({ page }) => {
+  const issue = await api.createIssue("Test Issue");
+  await page.goto(`/issues/${issue.id}`);
+});
+```
 
 ## Commit Rules
 
 - Use atomic commits grouped by logical intent.
-- Conventional format:
-  - `feat(scope): ...`
-  - `fix(scope): ...`
-  - `refactor(scope): ...`
-  - `docs: ...`
-  - `test(scope): ...`
-  - `chore(scope): ...`
-
-## CLI Release
-
-**Prerequisite:** A CLI release must accompany every Production deployment. When deploying to Production, always release a new CLI version as part of the process.
-
-1. Create a tag on the `main` branch: `git tag v0.x.x`
-2. Push the tag: `git push origin v0.x.x`
-3. GitHub Actions automatically triggers `release.yml`: runs Go tests → GoReleaser builds multi-platform binaries → publishes to GitHub Releases + Homebrew tap
-
-By default, bump the patch version each release (e.g. `v0.1.12` → `v0.1.13`), unless the user specifies a specific version.
+- Conventional format: `feat(scope)`, `fix(scope)`, `refactor(scope)`, `docs`, `test(scope)`, `chore(scope)`.
 
 ## Minimum Pre-Push Checks
 
@@ -218,7 +349,7 @@ Run verification only when the user explicitly asks for it.
 For targeted checks when requested:
 ```bash
 pnpm typecheck        # TypeScript type errors only
-pnpm test             # TS unit tests only (Vitest)
+pnpm test             # TS unit tests only (Vitest, all packages)
 make test             # Go tests only
 pnpm exec playwright test   # E2E only (requires backend + frontend running)
 ```
@@ -231,43 +362,29 @@ After writing or modifying code, always run the full verification pipeline:
 make check
 ```
 
-This runs all checks in sequence:
-1. TypeScript typecheck (`pnpm typecheck`)
-2. TypeScript unit tests (`pnpm test`)
-3. Go tests (`go test ./...`)
-4. E2E tests (auto-starts backend + frontend if needed, runs Playwright)
-
 **Workflow:**
 - Write code to satisfy the requirement
 - Run `make check`
-- If any step fails, read the error output, fix the code, and re-run `make check`
+- If any step fails, read the error output, fix the code, and re-run
 - Repeat until all checks pass
 - Only then consider the task complete
 
 **Quick iteration:** If you know only TypeScript or Go is affected, run individual checks first for faster feedback, then finish with a full `make check` before marking work complete.
 
-## E2E Test Patterns
+## CLI Release
 
-E2E tests should be self-contained. Use the `TestApiClient` fixture for data setup/teardown:
+**Prerequisite:** A CLI release must accompany every Production deployment.
 
-```typescript
-import { loginAsDefault, createTestApi } from "./helpers";
-import type { TestApiClient } from "./fixtures";
+1. Create a tag on the `main` branch: `git tag v0.x.x`
+2. Push the tag: `git push origin v0.x.x`
+3. GitHub Actions automatically triggers `release.yml`: runs Go tests → GoReleaser builds multi-platform binaries → publishes to GitHub Releases + Homebrew tap
 
-let api: TestApiClient;
+By default, bump the patch version each release (e.g. `v0.1.12` → `v0.1.13`), unless the user specifies a specific version.
 
-test.beforeEach(async ({ page }) => {
-  api = await createTestApi();       // logged-in API client
-  await loginAsDefault(page);        // browser session
-});
+## Multi-tenancy
 
-test.afterEach(async () => {
-  await api.cleanup();               // delete any data created during the test
-});
+All queries filter by `workspace_id`. Membership checks gate access. `X-Workspace-ID` header routes requests to the correct workspace.
 
-test("example", async ({ page }) => {
-  const issue = await api.createIssue("Test Issue");  // create via API
-  await page.goto(`/issues/${issue.id}`);             // test via UI
-  // api.cleanup() in afterEach removes the issue
-});
-```
+## Agent Assignees
+
+Assignees are polymorphic — can be a member or an agent. `assignee_type` + `assignee_id` on issues. Agents render with distinct styling (purple background, robot icon).

CLI_AND_DAEMON.md

@@ -7,8 +7,7 @@ The `multica` CLI connects your local machine to Multica. It handles authenticat
 ### Homebrew (macOS/Linux)
 
 ```bash
-brew tap multica-ai/tap
-brew install multica-cli
+brew install multica-ai/tap/multica
 ```
 
 ### Build from Source
@@ -22,14 +21,30 @@ cp server/bin/multica /usr/local/bin/multica
 
 ### Update
 
+```bash
+brew upgrade multica-ai/tap/multica
+```
+
+For install script or manual installs, use:
+
 ```bash
 multica update
 ```
 
-This auto-detects your installation method (Homebrew or manual) and upgrades accordingly.
+`multica update` auto-detects your installation method and upgrades accordingly.
 
 ## Quick Start
 
+```bash
+# One-command setup: configure, authenticate, and start the daemon
+multica setup
+
+# For self-hosted (local) deployments:
+multica setup self-host
+```
+
+Or step by step:
+
 ```bash
 # 1. Authenticate (opens browser for login)
 multica login
@@ -55,10 +70,10 @@ Opens your browser for OAuth authentication, creates a 90-day personal access to
 ### Token Login
 
 ```bash
-multica login --token
+multica login --token <mul_...>
 ```
 
-Authenticate by pasting a personal access token directly. Useful for headless environments.
+Authenticate using a personal access token directly. Useful for headless environments. Pass `--token=` with an empty value to be prompted interactively (so the token never lands in shell history).
 
 ### Check Status
 
@@ -125,6 +140,15 @@ The daemon auto-detects these AI CLIs on your PATH:
 |-----|---------|-------------|
 | [Claude Code](https://docs.anthropic.com/en/docs/claude-code) | `claude` | Anthropic's coding agent |
 | [Codex](https://github.com/openai/codex) | `codex` | OpenAI's coding agent |
+| [GitHub Copilot CLI](https://docs.github.com/en/copilot) | `copilot` | GitHub's coding agent (model routed by your GitHub entitlement) |
+| OpenCode | `opencode` | Open-source coding agent |
+| OpenClaw | `openclaw` | Open-source coding agent |
+| Hermes | `hermes` | Nous Research coding agent |
+| Gemini | `gemini` | Google's coding agent |
+| [Pi](https://pi.dev/) | `pi` | Pi coding agent |
+| [Cursor Agent](https://cursor.com/) | `cursor-agent` | Cursor's headless coding agent |
+| Kimi | `kimi` | Moonshot coding agent |
+| Kiro CLI | `kiro-cli` | Kiro ACP coding agent |
 
 You need at least one installed. The daemon registers each detected CLI as an available runtime.
 
@@ -145,11 +169,28 @@ Daemon behavior is configured via flags or environment variables:
 | Poll interval | `--poll-interval` | `MULTICA_DAEMON_POLL_INTERVAL` | `3s` |
 | Heartbeat interval | `--heartbeat-interval` | `MULTICA_DAEMON_HEARTBEAT_INTERVAL` | `15s` |
 | Agent timeout | `--agent-timeout` | `MULTICA_AGENT_TIMEOUT` | `2h` |
+| Codex semantic inactivity timeout | `--codex-semantic-inactivity-timeout` | `MULTICA_CODEX_SEMANTIC_INACTIVITY_TIMEOUT` | `10m` |
 | Max concurrent tasks | `--max-concurrent-tasks` | `MULTICA_DAEMON_MAX_CONCURRENT_TASKS` | `20` |
 | Daemon ID | `--daemon-id` | `MULTICA_DAEMON_ID` | hostname |
 | Device name | `--device-name` | `MULTICA_DAEMON_DEVICE_NAME` | hostname |
 | Runtime name | `--runtime-name` | `MULTICA_AGENT_RUNTIME_NAME` | `Local Agent` |
 | Workspaces root | — | `MULTICA_WORKSPACES_ROOT` | `~/multica_workspaces` |
+| GC enabled | — | `MULTICA_GC_ENABLED` | `true` (set `false`/`0` to disable) |
+| GC scan interval | — | `MULTICA_GC_INTERVAL` | `1h` |
+| GC TTL (done/cancelled issues) | — | `MULTICA_GC_TTL` | `24h` |
+| GC orphan TTL (no `.gc_meta.json`) | — | `MULTICA_GC_ORPHAN_TTL` | `72h` |
+| GC artifact TTL (open issues) | — | `MULTICA_GC_ARTIFACT_TTL` | `12h` (set `0` to disable) |
+| GC artifact patterns | — | `MULTICA_GC_ARTIFACT_PATTERNS` | `node_modules,.next,.turbo` |
+
+#### Workspace garbage collection
+
+The daemon periodically scans `MULTICA_WORKSPACES_ROOT` and reclaims disk space in three modes:
+
+- **Full task cleanup** — when an issue's status is `done` or `cancelled` and has been idle for `MULTICA_GC_TTL`, the entire task directory is removed.
+- **Orphan cleanup** — task directories with no `.gc_meta.json` (e.g. left over from a daemon crash) are removed once they exceed `MULTICA_GC_ORPHAN_TTL`.
+- **Artifact-only cleanup** — when a task has been completed for at least `MULTICA_GC_ARTIFACT_TTL` but the issue is still open, regenerable build outputs whose directory basename matches `MULTICA_GC_ARTIFACT_PATTERNS` are removed; the rest of the workdir (source, `.git`, `output/`, `logs/`, `.gc_meta.json`) is preserved so the agent can resume the same workdir on the next task.
+
+Patterns are basename-only — entries containing `/` or `\` are silently dropped — and `.git` subtrees are never descended into. The default list (`node_modules`, `.next`, `.turbo`) is intentionally narrow; extend it per deployment if your repos consistently produce other regenerable directories (for example, `MULTICA_GC_ARTIFACT_PATTERNS=node_modules,.next,.turbo,target,__pycache__`). To disable artifact cleanup entirely, set `MULTICA_GC_ARTIFACT_TTL=0`.
 
 Agent-specific overrides:
 
@@ -157,36 +198,68 @@ Agent-specific overrides:
 |----------|-------------|
 | `MULTICA_CLAUDE_PATH` | Custom path to the `claude` binary |
 | `MULTICA_CLAUDE_MODEL` | Override the Claude model used |
+| `MULTICA_CLAUDE_ARGS` | Default extra arguments for Claude Code runs |
 | `MULTICA_CODEX_PATH` | Custom path to the `codex` binary |
 | `MULTICA_CODEX_MODEL` | Override the Codex model used |
+| `MULTICA_CODEX_ARGS` | Default extra arguments for Codex runs |
+| `MULTICA_COPILOT_PATH` | Custom path to the `copilot` binary |
+| `MULTICA_COPILOT_MODEL` | Override the Copilot model used (note: GitHub Copilot routes models through your account entitlement, so this may not be honoured) |
+| `MULTICA_OPENCODE_PATH` | Custom path to the `opencode` binary |
+| `MULTICA_OPENCODE_MODEL` | Override the OpenCode model used |
+| `MULTICA_OPENCLAW_PATH` | Custom path to the `openclaw` binary |
+| `MULTICA_OPENCLAW_MODEL` | Override the OpenClaw model used |
+| `MULTICA_HERMES_PATH` | Custom path to the `hermes` binary |
+| `MULTICA_HERMES_MODEL` | Override the Hermes model used |
+| `MULTICA_GEMINI_PATH` | Custom path to the `gemini` binary |
+| `MULTICA_GEMINI_MODEL` | Override the Gemini model used |
+| `MULTICA_PI_PATH` | Custom path to the `pi` binary |
+| `MULTICA_PI_MODEL` | Override the Pi model used |
+| `MULTICA_CURSOR_PATH` | Custom path to the `cursor-agent` binary |
+| `MULTICA_CURSOR_MODEL` | Override the Cursor Agent model used |
+| `MULTICA_KIMI_PATH` | Custom path to the `kimi` binary |
+| `MULTICA_KIMI_MODEL` | Override the Kimi model used |
+| `MULTICA_KIRO_PATH` | Custom path to the `kiro-cli` binary |
+| `MULTICA_KIRO_MODEL` | Override the Kiro model used |
+
+`MULTICA_CLAUDE_ARGS` and `MULTICA_CODEX_ARGS` are parsed with POSIX shellword quoting, so values such as `--model "gpt-5.1 codex" --sandbox read-only` are split like a shell command line. Agent arguments are applied in this order: hardcoded Multica defaults, daemon-wide env defaults, then per-agent `custom_args` from the task.
 
 ### Self-Hosted Server
 
-When connecting to a self-hosted Multica instance, point the CLI to your server before logging in:
+When connecting to a self-hosted Multica instance, the easiest approach is:
 
 ```bash
-export MULTICA_APP_URL=https://app.example.com
-export MULTICA_SERVER_URL=wss://api.example.com/ws
+# One command — configures for localhost, authenticates, starts daemon
+multica setup self-host
+
+# Or for on-premise with custom domains:
+multica setup self-host --server-url https://api.example.com --app-url https://app.example.com
+```
+
+Or configure manually:
+
+```bash
+# Set URLs individually
+multica config set server_url http://localhost:8080
+multica config set app_url http://localhost:3000
+
+# For production with TLS:
+# multica config set server_url https://api.example.com
+# multica config set app_url https://app.example.com
 
 multica login
 multica daemon start
 ```
 
-Or set them persistently:
-
-```bash
-multica config set app_url https://app.example.com
-multica config set server_url wss://api.example.com/ws
-```
-
 ### Profiles
 
 Profiles let you run multiple daemons on the same machine — for example, one for production and one for a staging server.
 
 ```bash
-# Start a daemon for the staging server
-multica --profile staging login
-multica --profile staging daemon start
+# Set up a staging profile
+multica setup self-host --profile staging --server-url https://api-staging.example.com --app-url https://staging.example.com
+
+# Start its daemon
+multica daemon start --profile staging
 
 # Default profile runs separately
 multica daemon start
@@ -232,10 +305,12 @@ multica workspace members <workspace-id>
 multica issue list
 multica issue list --status in_progress
 multica issue list --priority urgent --assignee "Agent Name"
+multica issue list --assignee-id 5fb87ac7-23b5-4a7a-81fa-ed295a54545d
+multica issue list --full-id
 multica issue list --limit 20 --output json
 ```
 
-Available filters: `--status`, `--priority`, `--assignee`, `--limit`.
+Table output shows a routable issue `KEY` such as `MUL-123`; copy that key into follow-up commands like `issue get`, `issue comment list`, `issue status`, or `--parent`. Add `--full-id` when you need canonical UUIDs. Available filters: `--status`, `--priority`, `--assignee` / `--assignee-id`, `--project`, `--limit`. Use `--assignee-id <uuid>` for unambiguous filtering when names overlap.
 
 ### Get Issue
 
@@ -248,9 +323,10 @@ multica issue get <id> --output json
 
 ```bash
 multica issue create --title "Fix login bug" --description "..." --priority high --assignee "Lambda"
+multica issue create --title "Fix login bug" --assignee-id 5fb87ac7-23b5-4a7a-81fa-ed295a54545d
 ```
 
-Flags: `--title` (required), `--description`, `--status`, `--priority`, `--assignee`, `--parent`, `--due-date`.
+Flags: `--title` (required), `--description`, `--status`, `--priority`, `--assignee` / `--assignee-id`, `--parent`, `--project`, `--due-date`. Pass `--assignee-id <uuid>` (mutually exclusive with `--assignee`) when scripting against the IDs returned by `multica workspace members --output json` / `multica agent list --output json`.
 
 ### Update Issue
 
@@ -262,9 +338,12 @@ multica issue update <id> --title "New title" --priority urgent
 
 ```bash
 multica issue assign <id> --to "Lambda"
+multica issue assign <id> --to-id 5fb87ac7-23b5-4a7a-81fa-ed295a54545d
 multica issue assign <id> --unassign
 ```
 
+Pass `--to-id <uuid>` to assign by canonical UUID (mutually exclusive with `--to`); useful when names overlap across members and agents.
+
 ### Change Status
 
 ```bash
@@ -289,22 +368,127 @@ multica issue comment add <issue-id> --parent <comment-id> --content "Thanks!"
 multica issue comment delete <comment-id>
 ```
 
+### Subscribers
+
+```bash
+# List subscribers of an issue
+multica issue subscriber list <issue-id>
+
+# Subscribe yourself to an issue
+multica issue subscriber add <issue-id>
+
+# Subscribe another member or agent by name
+multica issue subscriber add <issue-id> --user "Lambda"
+
+# Unsubscribe yourself
+multica issue subscriber remove <issue-id>
+
+# Unsubscribe another member or agent
+multica issue subscriber remove <issue-id> --user "Lambda"
+```
+
+Subscribers receive notifications about issue activity (new comments, status changes, etc.). Without `--user`, the command acts on the caller.
+
 ### Execution History
 
 ```bash
 # List all execution runs for an issue
 multica issue runs <issue-id>
+multica issue runs <issue-id> --full-id
 multica issue runs <issue-id> --output json
 
 # View messages for a specific execution run
 multica issue run-messages <task-id>
+multica issue run-messages <short-task-id> --issue <issue-id>
 multica issue run-messages <task-id> --output json
 
 # Incremental fetch (only messages after a given sequence number)
 multica issue run-messages <task-id> --since 42 --output json
 ```
 
-The `runs` command shows all past and current executions for an issue, including running tasks. The `run-messages` command shows the detailed message log (tool calls, thinking, text, errors) for a single run. Use `--since` for efficient polling of in-progress runs.
+The `runs` command shows all past and current executions for an issue, including running tasks. Table output uses short task UUID prefixes by default; pass `--full-id` to print canonical task UUIDs. The `run-messages` command accepts full task UUIDs directly; copied short task prefixes must be scoped with `--issue <issue-id>` so the CLI only checks that issue's runs. It shows the detailed message log (tool calls, thinking, text, errors) for a single run. Use `--since` for efficient polling of in-progress runs.
+
+## Projects
+
+Projects group related issues (e.g. a sprint, an epic, a workstream). Every project
+belongs to a workspace and can optionally have a lead (member or agent).
+
+### List Projects
+
+```bash
+multica project list
+multica project list --status in_progress
+multica project list --output json
+```
+
+Available filters: `--status`.
+
+### Get Project
+
+```bash
+multica project get <id>
+multica project get <id> --output json
+```
+
+### Create Project
+
+```bash
+multica project create --title "2026 Week 16 Sprint" --icon "🏃" --lead "Lambda"
+```
+
+Flags: `--title` (required), `--description`, `--status`, `--icon`, `--lead`.
+
+### Update Project
+
+```bash
+multica project update <id> --title "New title" --status in_progress
+multica project update <id> --lead "Lambda"
+```
+
+Flags: `--title`, `--description`, `--status`, `--icon`, `--lead`.
+
+### Change Status
+
+```bash
+multica project status <id> in_progress
+```
+
+Valid statuses: `planned`, `in_progress`, `paused`, `completed`, `cancelled`.
+
+### Delete Project
+
+```bash
+multica project delete <id>
+```
+
+### Associating Issues with Projects
+
+Use the `--project` flag on `issue create` / `issue update` to attach an issue to a
+project, or on `issue list` to filter issues by project:
+
+```bash
+multica issue create --title "Login bug" --project <project-id>
+multica issue update <issue-id> --project <project-id>
+multica issue list --project <project-id>
+```
+
+## Setup
+
+```bash
+# One-command setup for Multica Cloud: configure, authenticate, and start the daemon
+multica setup
+
+# For local self-hosted deployments
+multica setup self-host
+
+# Custom ports
+multica setup self-host --port 9090 --frontend-port 4000
+
+# On-premise with custom domains
+multica setup self-host --server-url https://api.example.com --app-url https://app.example.com
+```
+
+`multica setup` configures the CLI, opens your browser for authentication, and starts the daemon — all in one step. Use `multica setup self-host` to connect to a self-hosted server instead of Multica Cloud.
 
 ## Configuration
 
@@ -319,11 +503,71 @@ Shows config file path, server URL, app URL, and default workspace.
 ### Set Values
 
 ```bash
-multica config set server_url wss://api.example.com/ws
+multica config set server_url https://api.example.com
 multica config set app_url https://app.example.com
 multica config set workspace_id <workspace-id>
 ```
 
+## Autopilot Commands
+
+Autopilots are scheduled/triggered automations that dispatch agent tasks (either by creating an issue or by running an agent directly).
+
+### List Autopilots
+
+```bash
+multica autopilot list
+multica autopilot list --full-id
+multica autopilot list --status active --output json
+```
+
+Autopilot table IDs are short UUID prefixes; follow-up autopilot commands accept copied prefixes when they are unique in the current workspace. Use `--full-id` to print canonical UUIDs.
+
+### Get Autopilot Details
+
+```bash
+multica autopilot get <id>
+multica autopilot get <id> --output json   # includes triggers
+```
+
+### Create / Update / Delete
+
+```bash
+multica autopilot create \
+  --title "Nightly bug triage" \
+  --description "Scan todo issues and prioritize." \
+  --agent "Lambda" \
+  --mode create_issue
+
+multica autopilot update <id> --status paused
+multica autopilot update <id> --description "New prompt"
+multica autopilot delete <id>
+```
+
+`--mode` currently only accepts `create_issue` (creates a new issue on each run and assigns it to the agent). The server data model also defines `run_only`, but the daemon task path doesn't yet resolve a workspace for runs without an issue, so it's not exposed by the CLI. `--agent` accepts either a name or UUID.
+
+### Manual Trigger
+
+```bash
+multica autopilot trigger <id>            # Fires the autopilot once, returns the run
+```
+
+### Run History
+
+```bash
+multica autopilot runs <id>
+multica autopilot runs <id> --limit 50 --output json
+```
+
+### Schedule Triggers
+
+```bash
+multica autopilot trigger-add <autopilot-id> --cron "0 9 * * 1-5" --timezone "America/New_York"
+multica autopilot trigger-update <autopilot-id> <trigger-id> --enabled=false
+multica autopilot trigger-delete <autopilot-id> <trigger-id>
+```
+
+Only cron-based `schedule` triggers are currently exposed via the CLI. The data model also defines `webhook` and `api` kinds, but there is no server endpoint that fires them yet, so they're not surfaced here.
+
 ## Other Commands
 
 ```bash

CLI_INSTALL.md

@@ -0,0 +1,201 @@
+# Multica CLI — Installation Guide for AI Agents
+
+> **Audience:** This document is designed for AI agents (Claude Code, Codex, etc.) to read and execute step-by-step. Each step includes the exact shell commands to run, expected outcomes, and error handling.
+
+## One-Line Instruction for Users
+
+Give this instruction to your AI agent:
+
+```
+Fetch https://github.com/multica-ai/multica/blob/main/CLI_INSTALL.md and follow the instructions to install Multica CLI, log in, and start the daemon on this machine.
+```
+
+---
+
+## Step 1: Check if Multica CLI is already installed
+
+Run:
+
+```bash
+multica version
+```
+
+- **If it prints a version string** (e.g. `multica v0.x.x`): skip to **Step 3**.
+- **If command not found**: continue to **Step 2**.
+
+---
+
+## Step 2: Install the Multica CLI
+
+> **Windows users:** Skip to [Option C: Windows (PowerShell)](#option-c-windows-powershell) below.
+
+### Option A: Homebrew (preferred — macOS/Linux)
+
+Check if Homebrew is available:
+
+```bash
+which brew
+```
+
+If `brew` is found, install via Homebrew:
+
+```bash
+brew install multica-ai/tap/multica
+```
+
+Then verify:
+
+```bash
+multica version
+```
+
+If the version prints successfully, skip to **Step 3**.
+
+To upgrade later, run:
+
+```bash
+brew upgrade multica-ai/tap/multica
+```
+
+### Option B: Download from GitHub Releases (macOS/Linux, no Homebrew)
+
+If Homebrew is not available, download the binary directly.
+
+Detect OS and architecture, then download the correct archive:
+
+```bash
+OS=$(uname -s | tr '[:upper:]' '[:lower:]')   # "darwin" or "linux"
+ARCH=$(uname -m)                                # "x86_64" or "arm64"
+
+# Normalize architecture name
+if [ "$ARCH" = "x86_64" ]; then
+  ARCH="amd64"
+fi
+
+# Get the latest release tag from GitHub
+LATEST=$(curl -sI https://github.com/multica-ai/multica/releases/latest | grep -i '^location:' | sed 's/.*tag\///' | tr -d '\r\n')
+
+# Download and extract
+VERSION="${LATEST#v}"
+curl -sL "https://github.com/multica-ai/multica/releases/download/${LATEST}/multica-cli-${VERSION}-${OS}-${ARCH}.tar.gz" -o /tmp/multica.tar.gz
+tar -xzf /tmp/multica.tar.gz -C /tmp multica
+sudo mv /tmp/multica /usr/local/bin/multica
+rm /tmp/multica.tar.gz
+```
+
+Verify:
+
+```bash
+multica version
+```
+
+**If this fails:**
+- Check that `/usr/local/bin` is in `$PATH`.
+- On Linux, you may need `chmod +x /usr/local/bin/multica`.
+- If `sudo` is not available, install to a user-writable directory: `mv /tmp/multica ~/.local/bin/multica` and ensure `~/.local/bin` is in `$PATH`.
+
+### Option C: Windows (PowerShell)
+
+Run in PowerShell (no admin required):
+
+```powershell
+irm https://raw.githubusercontent.com/multica-ai/multica/main/scripts/install.ps1 | iex
+```
+
+This downloads the latest Windows binary from GitHub Releases, installs it to `%USERPROFILE%\.multica\bin\`, and adds it to your user PATH.
+
+Verify:
+
+```powershell
+multica version
+```
+
+**If this fails:**
+- Restart your terminal so the updated PATH takes effect.
+- If you use Scoop, the installer will use it automatically: `scoop bucket add multica https://github.com/multica-ai/scoop-bucket.git && scoop install multica`
+- If your execution policy blocks the script: `Set-ExecutionPolicy -Scope CurrentUser -ExecutionPolicy RemoteSigned` then re-run.
+
+---
+
+## Step 3: Log in
+
+Run:
+
+```bash
+multica login
+```
+
+**Important:** This command opens a browser window for OAuth authentication. Tell the user:
+
+> "A browser window will open for Multica login. Please complete the authentication in your browser, then come back here."
+
+Wait for the command to complete. It will automatically discover and watch all workspaces the user belongs to.
+
+Verify:
+
+```bash
+multica auth status
+```
+
+Expected output should show the authenticated user and server URL.
+
+**If login fails:**
+- If no browser is available (headless environment), the user can generate a Personal Access Token at `https://app.multica.ai/settings` and run: `multica login --token <mul_...>` (use `--token=` with an empty value to be prompted interactively).
+- If the server URL needs to be customized: `multica config set server_url <url>` before logging in.
+
+---
+
+## Step 4: Start the daemon
+
+First, check if the daemon is already running:
+
+```bash
+multica daemon status
+```
+
+- **If status is "running"**: skip to **Step 5**.
+- **If status is "stopped"**: start it:
+
+```bash
+multica daemon start
+```
+
+Wait 3 seconds, then verify:
+
+```bash
+multica daemon status
+```
+
+Expected output should show `running` status with detected agents (e.g. `claude`, `codex`, `copilot`, `opencode`, `openclaw`, `hermes`, `gemini`, `pi`, `cursor-agent`).
+
+**If daemon fails to start:**
+- Check logs: `multica daemon logs`
+- If a port conflict occurs, the daemon may already be running under a different profile.
+- If no agents are detected, ensure at least one AI CLI (`claude`, `codex`, `copilot`, `opencode`, `openclaw`, `hermes`, `gemini`, `pi`, or `cursor-agent`) is installed and on the `$PATH`.
+
+---
+
+## Step 5: Verify everything is working
+
+Run:
+
+```bash
+multica daemon status
+```
+
+Confirm:
+1. Status is `running`
+2. At least one agent is listed (e.g. `claude`, `codex`, `copilot`, `opencode`, `openclaw`, `hermes`, `gemini`, `pi`, or `cursor-agent`)
+3. At least one workspace is being watched
+
+If the agents list is empty, tell the user:
+
+> "The Multica daemon is running but no AI agent CLIs were detected. Please install at least one supported CLI (`claude`, `codex`, `copilot`, `opencode`, `openclaw`, `hermes`, `gemini`, `pi`, or `cursor-agent`), then restart the daemon with `multica daemon stop && multica daemon start`."
+
+---
+
+## Summary
+
+When all steps are complete, inform the user:
+
+> "Multica CLI is installed and the daemon is running. Agents in your workspaces can now execute tasks on this machine. You can manage workspaces with `multica workspace list` and view daemon logs with `multica daemon logs -f`."

CONTRIBUTING.md

@@ -9,6 +9,7 @@ It covers:
 - isolated worktree development
 - the shared PostgreSQL model
 - testing and verification
+- full-stack isolated testing (backend + frontend + daemon from source)
 - troubleshooting and destructive reset options
 
 ## Development Model
@@ -94,59 +95,52 @@ FORCE=1 make worktree-env
 
 ## First-Time Setup
 
-### Main Checkout
+### Quick Start (recommended)
 
-From the main checkout:
+From any checkout (main or worktree):
+
+```bash
+make dev
+```
+
+This single command:
+
+- auto-detects whether you're in a main checkout or a worktree
+- creates the appropriate env file (`.env` or `.env.worktree`) if it doesn't exist
+- checks that prerequisites (Node.js, pnpm, Go, Docker) are installed
+- installs JavaScript dependencies
+- ensures the shared PostgreSQL container is running
+- creates the application database if it does not exist
+- runs all migrations
+- starts both backend and frontend
+
+### Explicit Setup (advanced)
+
+If you prefer separate control over setup and startup:
+
+#### Main Checkout
 
 ```bash
 cp .env.example .env
 make setup-main
-```
-
-What `make setup-main` does:
-
-- installs JavaScript dependencies with `pnpm install`
-- ensures the shared PostgreSQL container is running
-- creates the application database if it does not exist
-- runs all migrations against that database
-
-Start the app:
-
-```bash
 make start-main
 ```
 
-Stop the app processes:
+Stop:
 
 ```bash
 make stop-main
 ```
 
-This does not stop PostgreSQL.
-
-### Worktree
-
-From the worktree directory:
+#### Worktree
 
 ```bash
 make worktree-env
 make setup-worktree
-```
-
-What `make setup-worktree` does:
-
-- uses `.env.worktree`
-- ensures the shared PostgreSQL container is running
-- creates the worktree database if it does not exist
-- runs migrations against the worktree database
-
-Start the worktree app:
-
-```bash
 make start-worktree
 ```
 
-Stop the worktree app processes:
+Stop:
 
 ```bash
 make stop-worktree
@@ -171,17 +165,15 @@ Use a worktree when you want isolated data and separate app ports.
 ```bash
 git worktree add ../multica-feature -b feat/my-change main
 cd ../multica-feature
-make worktree-env
-make setup-worktree
-make start-worktree
+make dev
 ```
 
 After that, day-to-day commands are:
 
 ```bash
-make start-worktree
-make stop-worktree
-make check-worktree
+make dev              # start (re-runs setup if needed, idempotent)
+make stop-worktree    # stop
+make check-worktree   # verify
 ```
 
 ## Running Main and Worktree at the Same Time
@@ -317,6 +309,202 @@ make daemon
 The daemon authenticates using the CLI's stored token (`multica login`).
 It registers runtimes for all watched workspaces from the CLI config.
 
+## Full-Stack Isolated Testing
+
+This section covers running the complete stack (backend, frontend, daemon) from
+source in a fully isolated environment. Useful for testing end-to-end changes
+that span multiple components, or for automated CI/AI workflows that need zero
+human intervention.
+
+### Why Not Just `make daemon`?
+
+`make daemon` uses the system-installed CLI's stored token and connects to
+whatever server is configured in `~/.multica/config.json`. That's fine for
+day-to-day development against a shared server, but for fully isolated testing
+you need:
+
+- a local backend and frontend (from source)
+- a local daemon (from source) with its own profile
+- automated authentication (no browser login)
+- no interference with your production CLI config
+
+### Dynamic Profile Naming
+
+Each worktree must use a unique daemon profile to avoid collisions when
+multiple features run in parallel.
+
+The profile name is derived from the worktree directory using the same
+slug + hash pattern as `scripts/init-worktree-env.sh`:
+
+```bash
+WORKTREE_DIR="$(basename "$PWD")"
+SLUG="$(printf '%s' "$WORKTREE_DIR" | tr '[:upper:]' '[:lower:]' | sed 's/[^a-z0-9]/_/g; s/__*/_/g; s/^_//; s/_$//')"
+HASH="$(printf '%s' "$PWD" | cksum | awk '{print $1}')"
+OFFSET=$((HASH % 1000))
+PROFILE="dev-${SLUG}-${OFFSET}"
+```
+
+Example: worktree at `../multica-feat-auth` produces profile
+`dev-multica_feat_auth-347`, matching that worktree's port and database
+allocation.
+
+### Start the Isolated Environment
+
+Run all steps from the worktree root (where the Makefile is).
+
+#### 1. Start backend, frontend, and database
+
+```bash
+make dev
+```
+
+Wait for the backend to be healthy:
+
+```bash
+PORT=$(grep '^PORT=' .env.worktree 2>/dev/null || grep '^PORT=' .env | head -1 | cut -d= -f2)
+PORT=${PORT:-8080}
+SERVER="http://localhost:${PORT}"
+
+for i in $(seq 1 30); do
+  curl -sf "$SERVER/health" > /dev/null 2>&1 && break
+  sleep 2
+done
+```
+
+#### 2. Create a test user and token (automated auth)
+
+For deterministic local automation, set `MULTICA_DEV_VERIFICATION_CODE=888888`
+in your env file before starting the backend:
+
+```bash
+curl -s -X POST "$SERVER/auth/send-code" \
+  -H "Content-Type: application/json" \
+  -d '{"email": "dev@localhost"}'
+
+JWT=$(curl -s -X POST "$SERVER/auth/verify-code" \
+  -H "Content-Type: application/json" \
+  -d '{"email": "dev@localhost", "code": "888888"}' | jq -r '.token')
+
+PAT=$(curl -s -X POST "$SERVER/api/tokens" \
+  -H "Authorization: Bearer $JWT" \
+  -H "Content-Type: application/json" \
+  -d '{"name": "auto-dev", "expires_in_days": 365}' | jq -r '.token')
+```
+
+#### 3. Create a workspace
+
+```bash
+WS=$(curl -s -X POST "$SERVER/api/workspaces" \
+  -H "Authorization: Bearer $PAT" \
+  -H "Content-Type: application/json" \
+  -d '{"name": "Dev", "slug": "dev"}' | jq -r '.id')
+```
+
+#### 4. Compute profile name and write CLI config
+
+```bash
+# Compute profile (see Dynamic Profile Naming above)
+WORKTREE_DIR="$(basename "$PWD")"
+SLUG="$(printf '%s' "$WORKTREE_DIR" | tr '[:upper:]' '[:lower:]' | sed 's/[^a-z0-9]/_/g; s/__*/_/g; s/^_//; s/_$//')"
+HASH="$(printf '%s' "$PWD" | cksum | awk '{print $1}')"
+OFFSET=$((HASH % 1000))
+PROFILE="dev-${SLUG}-${OFFSET}"
+
+FRONTEND_PORT=$(grep '^FRONTEND_PORT=' .env.worktree 2>/dev/null || grep '^FRONTEND_PORT=' .env | head -1 | cut -d= -f2)
+FRONTEND_PORT=${FRONTEND_PORT:-3000}
+
+CONFIG_DIR="$HOME/.multica/profiles/$PROFILE"
+mkdir -p "$CONFIG_DIR"
+
+cat > "$CONFIG_DIR/config.json" << EOF
+{
+  "server_url": "$SERVER",
+  "app_url": "http://localhost:${FRONTEND_PORT}",
+  "token": "$PAT",
+  "workspace_id": "$WS",
+  "watched_workspaces": [{"id": "$WS", "name": "Dev"}]
+}
+EOF
+```
+
+#### 5. Start the daemon from source
+
+```bash
+make cli ARGS="daemon start --profile $PROFILE"
+```
+
+The daemon runs from the current worktree's Go source, connecting to the
+local backend. Agent-executed `multica` commands automatically use the same
+binary (the daemon prepends its own directory to `PATH`).
+
+### Stop the Isolated Environment
+
+```bash
+# Compute profile (same formula)
+PROFILE="dev-$(printf '%s' "$(basename "$PWD")" | tr '[:upper:]' '[:lower:]' | sed 's/[^a-z0-9]/_/g; s/__*/_/g; s/^_//; s/_$//')-$(( $(printf '%s' "$PWD" | cksum | awk '{print $1}') % 1000 ))"
+
+# 1. Stop daemon
+make cli ARGS="daemon stop --profile $PROFILE"
+
+# 2. Stop backend + frontend
+make stop            # main checkout
+make stop-worktree   # worktree checkout
+
+# 3. (Optional) Stop shared PostgreSQL
+make db-down
+
+# 4. (Optional) Clean build artifacts
+make clean
+
+# 5. (Optional) Remove profile config
+rm -rf "$HOME/.multica/profiles/$PROFILE"
+```
+
+### Desktop App Local Testing
+
+To test the Electron desktop app against a local backend:
+
+```bash
+# After backend is running (make dev)
+pnpm dev:desktop
+```
+
+This automatically:
+
+1. Compiles the `multica` CLI from `server/cmd/multica` into
+   `apps/desktop/resources/bin/multica`
+2. Creates an isolated profile named `desktop-localhost-<PORT>`
+3. Starts and manages its own daemon instance
+4. Connects to the local backend
+
+Login in the Desktop UI with `dev@localhost` and the generated code from the
+backend logs. If you set `MULTICA_DEV_VERIFICATION_CODE=888888` before starting
+the backend, you can use `888888` instead.
+
+If the backend runs on a non-default port (worktree), create
+`apps/desktop/.env.development.local`:
+
+```bash
+VITE_API_URL=http://localhost:<backend-port>
+VITE_WS_URL=ws://localhost:<backend-port>/ws
+```
+
+### Isolation Guarantee
+
+Nothing in this flow touches the system-installed `multica` or the default
+`~/.multica/config.json`:
+
+| Resource | System / Production | Local Dev (per-worktree) |
+|---|---|---|
+| Config | `~/.multica/config.json` | `~/.multica/profiles/dev-<slug>-<hash>/config.json` |
+| Daemon PID | `~/.multica/daemon.pid` | `~/.multica/profiles/dev-<slug>-<hash>/daemon.pid` |
+| Health port | `19514` | `19514 + 1 + (name_hash % 1000)` |
+| Workspaces dir | `~/multica_workspaces/` | `~/multica_workspaces_dev-<slug>-<hash>/` |
+| Database | remote / production | local Docker: `multica_<slug>_<hash>` |
+| Desktop profile | `desktop-api.multica.ai` | `desktop-localhost-<port>` |
+
+Multiple worktrees can run simultaneously without conflict.
+
 ## Troubleshooting
 
 ### Missing Env File
@@ -407,6 +595,19 @@ If you want to stop PostgreSQL and keep your local databases:
 make db-down
 ```
 
+If you want a fresh database for the current checkout only (drops the
+database named in `POSTGRES_DB`, recreates it, and runs all migrations):
+
+```bash
+make stop        # stop backend/frontend first
+make db-reset
+make start
+```
+
+- only affects the current env's database; other worktree databases are untouched
+- refuses to run if `DATABASE_URL` points at a remote host
+- pass `ENV_FILE=.env.worktree` to target a specific worktree
+
 If you want to wipe all local PostgreSQL data for this repo:
 
 ```bash
@@ -424,9 +625,7 @@ Warning:
 ### Stable Main Environment
 
 ```bash
-cp .env.example .env
-make setup-main
-make start-main
+make dev
 ```
 
 ### Feature Worktree
@@ -434,9 +633,7 @@ make start-main
 ```bash
 git worktree add ../multica-feature -b feat/my-change main
 cd ../multica-feature
-make worktree-env
-make setup-worktree
-make start-worktree
+make dev
 ```
 
 ### Return to a Previously Configured Worktree

Dockerfile

@@ -15,7 +15,7 @@ COPY server/ ./server/
 # Build binaries
 ARG VERSION=dev
 ARG COMMIT=unknown
-RUN cd server && CGO_ENABLED=0 go build -ldflags "-s -w" -o bin/server ./cmd/server
+RUN cd server && CGO_ENABLED=0 go build -ldflags "-s -w -X main.version=${VERSION} -X main.commit=${COMMIT}" -o bin/server ./cmd/server
 RUN cd server && CGO_ENABLED=0 go build -ldflags "-s -w -X main.version=${VERSION} -X main.commit=${COMMIT}" -o bin/multica ./cmd/multica
 RUN cd server && CGO_ENABLED=0 go build -ldflags "-s -w" -o bin/migrate ./cmd/migrate
 
@@ -30,7 +30,9 @@ COPY --from=builder /src/server/bin/server .
 COPY --from=builder /src/server/bin/multica .
 COPY --from=builder /src/server/bin/migrate .
 COPY server/migrations/ ./migrations/
+COPY docker/entrypoint.sh .
+RUN sed -i 's/\r$//' entrypoint.sh && chmod +x entrypoint.sh
 
 EXPOSE 8080
 
-ENTRYPOINT ["./server"]
+ENTRYPOINT ["./entrypoint.sh"]

Dockerfile.web

@@ -0,0 +1,72 @@
+# --- Dependencies ---
+FROM node:22-alpine AS deps
+
+RUN corepack enable && corepack prepare pnpm@10.28.2 --activate
+
+WORKDIR /app
+
+# Copy workspace config and all package.json files for dependency resolution
+COPY pnpm-lock.yaml pnpm-workspace.yaml package.json turbo.json .npmrc ./
+COPY apps/web/package.json apps/web/
+COPY packages/core/package.json packages/core/
+COPY packages/ui/package.json packages/ui/
+COPY packages/views/package.json packages/views/
+COPY packages/tsconfig/package.json packages/tsconfig/
+COPY packages/eslint-config/package.json packages/eslint-config/
+
+RUN pnpm install --frozen-lockfile
+
+# --- Build ---
+FROM node:22-alpine AS builder
+
+RUN corepack enable && corepack prepare pnpm@10.28.2 --activate
+
+WORKDIR /app
+
+# Copy installed dependencies (preserves pnpm symlink structure)
+COPY --from=deps /app ./
+
+# Copy source
+COPY package.json turbo.json pnpm-workspace.yaml ./
+COPY apps/web/ apps/web/
+COPY packages/ packages/
+
+# Re-link after source overlay (fixes any symlinks overwritten by COPY)
+RUN pnpm install --frozen-lockfile --offline
+
+# Set build-time env: tells Next.js rewrites to proxy API calls to the backend service
+ARG REMOTE_API_URL=http://backend:8080
+ARG NEXT_PUBLIC_WS_URL
+ARG NEXT_PUBLIC_APP_VERSION=dev
+ENV REMOTE_API_URL=$REMOTE_API_URL
+ENV NEXT_PUBLIC_WS_URL=$NEXT_PUBLIC_WS_URL
+ENV NEXT_PUBLIC_APP_VERSION=$NEXT_PUBLIC_APP_VERSION
+ENV STANDALONE=true
+
+# Build the web app (standalone output for minimal runtime)
+RUN pnpm --filter @multica/web build
+
+# --- Runtime ---
+FROM node:22-alpine AS runner
+
+WORKDIR /app
+
+ENV NODE_ENV=production
+
+RUN addgroup --system --gid 1001 nodejs && \
+    adduser --system --uid 1001 nextjs
+
+# Copy standalone output (includes traced node_modules)
+COPY --from=builder --chown=nextjs:nodejs /app/apps/web/.next/standalone ./
+# Copy static files (not included in standalone)
+COPY --from=builder --chown=nextjs:nodejs /app/apps/web/.next/static ./apps/web/.next/static
+# Copy public assets
+COPY --from=builder --chown=nextjs:nodejs /app/apps/web/public ./apps/web/public
+
+USER nextjs
+
+EXPOSE 3000
+ENV PORT=3000
+ENV HOSTNAME=0.0.0.0
+
+CMD ["node", "apps/web/server.js"]

LICENSE

@@ -1,199 +1,44 @@
-                                 Apache License
-                           Version 2.0, January 2004
-                        http://www.apache.org/licenses/
+# Open Source License
 
-   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+Multica is licensed under a modified version of the Apache License 2.0, with the following additional conditions:
 
-   1. Definitions.
+1. Multica may be utilized commercially, including as a backend service for
+   other applications or as a task management platform for enterprises.
+   Should the conditions below be met, a commercial license must be obtained
+   from the producer:
 
-      "License" shall mean the terms and conditions for use, reproduction,
-      and distribution as defined by Sections 1 through 9 of this document.
+   a. Hosted or embedded service: Unless explicitly authorized by Multica
+      in writing, you may not use the Multica source code to provide a
+      hosted service to third parties, or embed Multica as a component of
+      a product or service that is sold, licensed, or otherwise
+      commercially distributed to third parties.
 
-      "Licensor" shall mean the copyright owner or entity authorized by
-      the copyright owner that is granting the License.
+      - This restriction applies to offering Multica (in whole or
+        substantial part) as a SaaS platform, a managed service, or as
+        an integrated component within another commercial offering.
+      - Internal use within a single organization (including multiple
+        workspaces) does not require a commercial license.
 
-      "Legal Entity" shall mean the union of the acting entity and all
-      other entities that control, are controlled by, or are under common
-      control with that entity. For the purposes of this definition,
-      "control" means (i) the power, direct or indirect, to cause the
-      direction or management of such entity, whether by contract or
-      otherwise, or (ii) ownership of fifty percent (50%) or more of the
-      outstanding shares, or (iii) beneficial ownership of such entity.
+   b. LOGO and copyright information: In the process of using Multica's
+      frontend, you may not remove or modify the LOGO or copyright
+      information in the Multica console or applications. This restriction
+      is inapplicable to uses of Multica that do not involve its frontend.
 
-      "You" (or "Your") shall mean an individual or Legal Entity
-      exercising permissions granted by this License.
+      - Frontend Definition: For the purposes of this license, the
+        "frontend" of Multica includes all components located in the
+        `apps/web/` directory when running Multica from the raw source
+        code, or the "web" image when running Multica with Docker.
 
-      "Source" form shall mean the preferred form for making modifications,
-      including but not limited to software source code, documentation
-      source, and configuration files.
+2. As a contributor, you should agree that:
 
-      "Object" form shall mean any form resulting from mechanical
-      transformation or translation of a Source form, including but
-      not limited to compiled object code, generated documentation,
-      and conversions to other media types.
+   a. The producer can adjust the open-source agreement to be more strict
+      or relaxed as deemed necessary.
 
-      "Work" shall mean the work of authorship, whether in Source or
-      Object form, made available under the License, as indicated by a
-      copyright notice that is included in or attached to the work
-      (an example is provided in the Appendix below).
+   b. Your contributed code may be used for commercial purposes, including
+      but not limited to its cloud business operations.
 
-      "Derivative Works" shall mean any work, whether in Source or Object
-      form, that is based on (or derived from) the Work and for which the
-      editorial revisions, annotations, elaborations, or other modifications
-      represent, as a whole, an original work of authorship. For the purposes
-      of this License, Derivative Works shall not include works that remain
-      separable from, or merely link (or bind by name) to the interfaces of,
-      the Work and Derivative Works thereof.
+Apart from the specific conditions mentioned above, all other rights and
+restrictions follow the Apache License 2.0. Detailed information about the
+Apache License 2.0 can be found at http://www.apache.org/licenses/LICENSE-2.0.
 
-      "Contribution" shall mean any work of authorship, including
-      the original version of the Work and any modifications or additions
-      to that Work or Derivative Works thereof, that is intentionally
-      submitted to the Licensor for inclusion in the Work by the copyright owner
-      or by an individual or Legal Entity authorized to submit on behalf of
-      the copyright owner. For the purposes of this definition, "submitted"
-      means any form of electronic, verbal, or written communication sent
-      to the Licensor or its representatives, including but not limited to
-      communication on electronic mailing lists, source code control systems,
-      and issue tracking systems that are managed by, or on behalf of, the
-      Licensor for the purpose of discussing and improving the Work, but
-      excluding communication that is conspicuously marked or otherwise
-      designated in writing by the copyright owner as "Not a Contribution."
-
-      "Contributor" shall mean Licensor and any individual or Legal Entity
-      on behalf of whom a Contribution has been received by the Licensor and
-      subsequently incorporated within the Work.
-
-   2. Grant of Copyright License. Subject to the terms and conditions of
-      this License, each Contributor hereby grants to You a perpetual,
-      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
-      copyright license to reproduce, prepare Derivative Works of,
-      publicly display, publicly perform, sublicense, and distribute the
-      Work and such Derivative Works in Source or Object form.
-
-   3. Grant of Patent License. Subject to the terms and conditions of
-      this License, each Contributor hereby grants to You a perpetual,
-      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
-      (except as stated in this section) patent license to make, have made,
-      use, offer to sell, sell, import, and otherwise transfer the Work,
-      where such license applies only to those patent claims licensable
-      by such Contributor that are necessarily infringed by their
-      Contribution(s) alone or by combination of their Contribution(s)
-      with the Work to which such Contribution(s) was submitted. If You
-      institute patent litigation against any entity (including a
-      cross-claim or counterclaim in a lawsuit) alleging that the Work
-      or a Contribution incorporated within the Work constitutes direct
-      or contributory patent infringement, then any patent licenses
-      granted to You under this License for that Work shall terminate
-      as of the date such litigation is filed.
-
-   4. Redistribution. You may reproduce and distribute copies of the
-      Work or Derivative Works thereof in any medium, with or without
-      modifications, and in Source or Object form, provided that You
-      meet the following conditions:
-
-      (a) You must give any other recipients of the Work or
-          Derivative Works a copy of this License; and
-
-      (b) You must cause any modified files to carry prominent notices
-          stating that You changed the files; and
-
-      (c) You must retain, in the Source form of any Derivative Works
-          that You distribute, all copyright, patent, trademark, and
-          attribution notices from the Source form of the Work,
-          excluding those notices that do not pertain to any part of
-          the Derivative Works; and
-
-      (d) If the Work includes a "NOTICE" text file as part of its
-          distribution, then any Derivative Works that You distribute must
-          include a readable copy of the attribution notices contained
-          within such NOTICE file, excluding any notices that do not
-          pertain to any part of the Derivative Works, in at least one
-          of the following places: within a NOTICE text file distributed
-          as part of the Derivative Works; within the Source form or
-          documentation, if provided along with the Derivative Works; or,
-          within a display generated by the Derivative Works, if and
-          wherever such third-party notices normally appear. The contents
-          of the NOTICE file are for informational purposes only and
-          do not modify the License. You may add Your own attribution
-          notices within Derivative Works that You distribute, alongside
-          or as an addendum to the NOTICE text from the Work, provided
-          that such additional attribution notices cannot be construed
-          as modifying the License.
-
-      You may add Your own copyright statement to Your modifications and
-      may provide additional or different license terms and conditions
-      for use, reproduction, or distribution of Your modifications, or
-      for any such Derivative Works as a whole, provided Your use,
-      reproduction, and distribution of the Work otherwise complies with
-      the conditions stated in this License.
-
-   5. Submission of Contributions. Unless You explicitly state otherwise,
-      any Contribution intentionally submitted for inclusion in the Work
-      by You to the Licensor shall be under the terms and conditions of
-      this License, without any additional terms or conditions.
-      Notwithstanding the above, nothing herein shall supersede or modify
-      the terms of any separate license agreement you may have executed
-      with Licensor regarding such Contributions.
-
-   6. Trademarks. This License does not grant permission to use the trade
-      names, trademarks, service marks, or product names of the Licensor,
-      except as required for reasonable and customary use in describing the
-      origin of the Work and reproducing the content of the NOTICE file.
-
-   7. Disclaimer of Warranty. Unless required by applicable law or
-      agreed to in writing, Licensor provides the Work (and each
-      Contributor provides its Contributions) on an "AS IS" BASIS,
-      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
-      implied, including, without limitation, any warranties or conditions
-      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
-      PARTICULAR PURPOSE. You are solely responsible for determining the
-      appropriateness of using or redistributing the Work and assume any
-      risks associated with Your exercise of permissions under this License.
-
-   8. Limitation of Liability. In no event and under no legal theory,
-      whether in tort (including negligence), contract, or otherwise,
-      unless required by applicable law (such as deliberate and grossly
-      negligent acts) or agreed to in writing, shall any Contributor be
-      liable to You for damages, including any direct, indirect, special,
-      incidental, or consequential damages of any character arising as a
-      result of this License or out of the use or inability to use the
-      Work (including but not limited to damages for loss of goodwill,
-      work stoppage, computer failure or malfunction, or any and all
-      other commercial damages or losses), even if such Contributor
-      has been advised of the possibility of such damages.
-
-   9. Accepting Warranty or Additional Liability. While redistributing
-      the Work or Derivative Works thereof, You may choose to offer,
-      and charge a fee for, acceptance of support, warranty, indemnity,
-      or other liability obligations and/or rights consistent with this
-      License. However, in accepting such obligations, You may act only
-      on Your own behalf and on Your sole responsibility, not on behalf
-      of any other Contributor, and only if You agree to indemnify,
-      defend, and hold each Contributor harmless for any liability
-      incurred by, or claims asserted against, such Contributor by reason
-      of your accepting any such warranty or additional liability.
-
-   END OF TERMS AND CONDITIONS
-
-   APPENDIX: How to apply the Apache License to your work.
-
-      To apply the Apache License to your work, attach the following
-      boilerplate notice, with the fields enclosed by brackets "[]"
-      replaced with your own identifying information. (Don't include
-      the brackets!)  The text should be enclosed in the appropriate
-      comment syntax for the file format. Please also get an
-      "Implied Patent License" from your patent counsel.
-
-   Copyright 2025 Multica
-
-   Licensed under the Apache License, Version 2.0 (the "License");
-   you may not use this file except in compliance with the License.
-   You may obtain a copy of the License at
-
-       http://www.apache.org/licenses/LICENSE-2.0
-
-   Unless required by applicable law or agreed to in writing, software
-   distributed under the License is distributed on an "AS IS" BASIS,
-   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-   See the License for the specific language governing permissions and
-   limitations under the License.
+© 2025 Multica, Inc.

Makefile

@@ -1,4 +1,4 @@
-.PHONY: dev daemon cli multica build test migrate-up migrate-down sqlc seed clean setup start stop check worktree-env setup-main start-main stop-main check-main setup-worktree start-worktree stop-worktree check-worktree db-up db-down
+.PHONY: help makehelp dev server daemon cli multica build test migrate-up migrate-down sqlc seed clean setup start stop check worktree-env setup-main start-main stop-main check-main setup-worktree start-worktree stop-worktree check-worktree db-up db-down db-reset selfhost selfhost-build selfhost-stop
 
 MAIN_ENV_FILE ?= .env
 WORKTREE_ENV_FILE ?= .env.worktree
@@ -36,10 +36,123 @@ define REQUIRE_ENV
 	fi
 endef
 
-# ---------- One-click commands ----------
+# Default target changed from selfhost to help: bare `make` now prints this help
+# instead of launching a full Docker Compose build, which is safer for onboarding.
+.DEFAULT_GOAL := help
 
-# First-time setup: install deps, start DB, run migrations
-setup:
+##@ Help
+
+help: ## Show available make targets and common local workflows
+	@awk 'BEGIN {FS = ":.*## "; printf "\nUsage:\n  make \033[36m<target>\033[0m\n\nQuick start:\n  \033[36mmake dev\033[0m          Bootstrap the current checkout and start everything\n  \033[36mmake check\033[0m        Run the full local verification pipeline\n\nCheckout modes:\n  Main checkout uses \033[36m.env\033[0m\n  Worktrees use \033[36m.env.worktree\033[0m (generate with \033[36mmake worktree-env\033[0m)\n\n"} \
+		/^##@/ {printf "\n\033[1m%s\033[0m\n", substr($$0, 5); next} \
+		/^[a-zA-Z0-9_.-]+:.*## / {printf "  \033[36m%-18s\033[0m %s\n", $$1, $$2}' $(MAKEFILE_LIST)
+
+makehelp: help ## Alias for `make help`
+
+# ---------- Self-hosting (Docker Compose) ----------
+##@ Self-hosting
+
+selfhost: ## Create .env if needed, then pull and start the official self-hosted images
+	@if [ ! -f .env ]; then \
+		echo "==> Creating .env from .env.example..."; \
+		cp .env.example .env; \
+		JWT=$$(openssl rand -hex 32); \
+		if [ "$$(uname)" = "Darwin" ]; then \
+			sed -i '' "s/^JWT_SECRET=.*/JWT_SECRET=$$JWT/" .env; \
+		else \
+			sed -i "s/^JWT_SECRET=.*/JWT_SECRET=$$JWT/" .env; \
+		fi; \
+		echo "==> Generated random JWT_SECRET"; \
+	fi
+	@echo "==> Pulling official Multica images..."
+	@if ! docker compose -f docker-compose.selfhost.yml pull; then \
+		echo ""; \
+		echo "Official images for tag '$${MULTICA_IMAGE_TAG:-latest}' are not published yet."; \
+		echo "If this is before the first GHCR release, build from the current checkout:"; \
+		echo "  make selfhost-build"; \
+		exit 1; \
+	fi
+	@echo "==> Starting Multica via Docker Compose..."
+	docker compose -f docker-compose.selfhost.yml up -d
+	@echo "==> Waiting for backend to be ready..."
+	@for i in $$(seq 1 30); do \
+		if curl -sf http://localhost:$${PORT:-8080}/health > /dev/null 2>&1; then \
+			break; \
+		fi; \
+		sleep 2; \
+	done
+	@if curl -sf http://localhost:$${PORT:-8080}/health > /dev/null 2>&1; then \
+		echo ""; \
+		echo "✓ Multica is running!"; \
+		echo "  Frontend: http://localhost:$${FRONTEND_PORT:-3000}"; \
+		echo "  Backend:  http://localhost:$${PORT:-8080}"; \
+		echo ""; \
+		echo "Images: $${MULTICA_BACKEND_IMAGE:-ghcr.io/multica-ai/multica-backend}:$${MULTICA_IMAGE_TAG:-latest}"; \
+		echo "        $${MULTICA_WEB_IMAGE:-ghcr.io/multica-ai/multica-web}:$${MULTICA_IMAGE_TAG:-latest}"; \
+		echo ""; \
+		echo "Log in: configure RESEND_API_KEY in .env for email codes,"; \
+		echo "        or read the generated code from backend logs when Resend is unset."; \
+		echo ""; \
+		echo "Next — install the CLI and connect your machine:"; \
+		echo "  brew install multica-ai/tap/multica"; \
+		echo "  multica setup self-host"; \
+	else \
+		echo ""; \
+		echo "Services are still starting. Check logs:"; \
+		echo "  docker compose -f docker-compose.selfhost.yml logs"; \
+	fi
+
+selfhost-build: ## Build backend/web from the current checkout and start the self-hosted stack
+	@if [ ! -f .env ]; then \
+		echo "==> Creating .env from .env.example..."; \
+		cp .env.example .env; \
+		JWT=$$(openssl rand -hex 32); \
+		if [ "$$(uname)" = "Darwin" ]; then \
+			sed -i '' "s/^JWT_SECRET=.*/JWT_SECRET=$$JWT/" .env; \
+		else \
+			sed -i "s/^JWT_SECRET=.*/JWT_SECRET=$$JWT/" .env; \
+		fi; \
+		echo "==> Generated random JWT_SECRET"; \
+	fi
+	@echo "==> Building Multica from the current checkout..."
+	docker compose -f docker-compose.selfhost.yml -f docker-compose.selfhost.build.yml up -d --build
+	@echo "==> Waiting for backend to be ready..."
+	@for i in $$(seq 1 30); do \
+		if curl -sf http://localhost:$${PORT:-8080}/health > /dev/null 2>&1; then \
+			break; \
+		fi; \
+		sleep 2; \
+	done
+	@if curl -sf http://localhost:$${PORT:-8080}/health > /dev/null 2>&1; then \
+		echo ""; \
+		echo "✓ Multica is running!"; \
+		echo "  Frontend: http://localhost:$${FRONTEND_PORT:-3000}"; \
+		echo "  Backend:  http://localhost:$${PORT:-8080}"; \
+		echo ""; \
+		echo "Log in: configure RESEND_API_KEY in .env for email codes,"; \
+		echo "        or read the generated code from backend logs when Resend is unset."; \
+		echo ""; \
+		echo "Built images locally via docker-compose.selfhost.build.yml."; \
+		echo "Local tags: multica-backend:dev and multica-web:dev."; \
+		echo ""; \
+		echo "Next — install the CLI and connect your machine:"; \
+		echo "  brew install multica-ai/tap/multica"; \
+		echo "  multica setup self-host"; \
+	else \
+		echo ""; \
+		echo "Services are still starting. Check logs:"; \
+		echo "  docker compose -f docker-compose.selfhost.yml logs"; \
+	fi
+
+selfhost-stop: ## Stop the self-hosted Docker Compose stack
+	@echo "==> Stopping Multica services..."
+	docker compose -f docker-compose.selfhost.yml down
+	@echo "✓ All services stopped."
+
+# ---------- One-click commands ----------
+##@ One-click
+
+setup: ## Prepare the current checkout from its env file: install deps, ensure DB, run migrations
 	$(REQUIRE_ENV)
 	@echo "==> Using env file: $(ENV_FILE)"
 	@echo "==> Installing dependencies..."
@@ -50,110 +163,148 @@ setup:
 	@echo ""
 	@echo "✓ Setup complete! Run 'make start' to launch the app."
 
-# Start all services (backend + frontend)
-start:
+start: ## Start backend and frontend for the current checkout and run migrations first
 	$(REQUIRE_ENV)
 	@echo "Using env file: $(ENV_FILE)"
 	@echo "Backend: http://localhost:$(PORT)"
 	@echo "Frontend: http://localhost:$(FRONTEND_PORT)"
 	@bash scripts/ensure-postgres.sh "$(ENV_FILE)"
+	@echo "Running migrations..."
+	cd server && go run ./cmd/migrate up
 	@echo "Starting backend and frontend..."
 	@trap 'kill 0' EXIT; \
 		(cd server && go run ./cmd/server) & \
 		pnpm dev:web & \
 		wait
 
-# Stop all services
-stop:
+stop: ## Stop backend and frontend processes for the current checkout
 	$(REQUIRE_ENV)
 	@echo "Stopping services..."
 	@-lsof -ti:$(PORT) | xargs kill -9 2>/dev/null
 	@-lsof -ti:$(FRONTEND_PORT) | xargs kill -9 2>/dev/null
-	@echo "✓ App processes stopped. Shared PostgreSQL is still running on localhost:5432."
+	@case "$(DATABASE_URL)" in \
+		""|*@localhost:*|*@localhost/*|*@127.0.0.1:*|*@127.0.0.1/*|*@\[::1\]:*|*@\[::1\]/*) \
+			echo "✓ App processes stopped. Shared PostgreSQL is still running on localhost:$(POSTGRES_PORT)." ;; \
+		*) \
+			echo "✓ App processes stopped. Remote PostgreSQL was not affected." ;; \
+	esac
 
-# Full verification: typecheck + unit tests + Go tests + E2E
-check:
+check: ## Run typecheck, TS tests, Go tests, and Playwright E2E for the current checkout
 	$(REQUIRE_ENV)
 	@ENV_FILE="$(ENV_FILE)" bash scripts/check.sh
 
-db-up:
+db-up: ## Start the shared PostgreSQL container used by main and worktrees
 	@$(COMPOSE) up -d postgres
 
-db-down:
+db-down: ## Stop the shared PostgreSQL container without removing its Docker volume
 	@$(COMPOSE) down
 
-worktree-env:
+# Drop + recreate the current env's database, then run all migrations.
+# Use for a clean slate in local dev. Only affects the DB named in
+# ENV_FILE (POSTGRES_DB); the shared postgres container and other
+# worktree DBs are untouched. Refuses to run against a remote host.
+db-reset: ## Drop and recreate the current env's database, then re-run all migrations
+	$(REQUIRE_ENV)
+	@case "$(DATABASE_URL)" in \
+		""|*@localhost:*|*@localhost/*|*@127.0.0.1:*|*@127.0.0.1/*|*@\[::1\]:*|*@\[::1\]/*) ;; \
+		*) echo "Refusing to reset: DATABASE_URL points at a remote host."; exit 1 ;; \
+	esac
+	@bash scripts/ensure-postgres.sh "$(ENV_FILE)"
+	@echo "==> Dropping and recreating database '$(POSTGRES_DB)'..."
+	@$(COMPOSE) exec -T postgres psql -U $(POSTGRES_USER) -d postgres -v ON_ERROR_STOP=1 \
+		-c "DROP DATABASE IF EXISTS \"$(POSTGRES_DB)\" WITH (FORCE);" \
+		-c "CREATE DATABASE \"$(POSTGRES_DB)\";"
+	@echo "==> Running migrations..."
+	cd server && go run ./cmd/migrate up
+	@echo ""
+	@echo "✓ Database '$(POSTGRES_DB)' reset. Run 'make start' to launch the app."
+
+worktree-env: ## Generate .env.worktree with a unique DB name and app ports for this worktree
 	@bash scripts/init-worktree-env.sh .env.worktree
 
-setup-main:
+setup-main: ## Prepare the main checkout using .env
 	@$(MAKE) setup ENV_FILE=$(MAIN_ENV_FILE)
 
-start-main:
+start-main: ## Start the main checkout using .env
 	@$(MAKE) start ENV_FILE=$(MAIN_ENV_FILE)
 
-stop-main:
+stop-main: ## Stop the main checkout processes defined by .env
 	@$(MAKE) stop ENV_FILE=$(MAIN_ENV_FILE)
 
-check-main:
+check-main: ## Run the full verification pipeline for the main checkout
 	@ENV_FILE=$(MAIN_ENV_FILE) bash scripts/check.sh
 
-setup-worktree:
-	@echo "==> Generating $(WORKTREE_ENV_FILE) with unique ports..."
-	@FORCE=1 bash scripts/init-worktree-env.sh $(WORKTREE_ENV_FILE)
+setup-worktree: ## Ensure .env.worktree exists, then prepare this worktree
+	@if [ ! -f "$(WORKTREE_ENV_FILE)" ]; then \
+		echo "==> Generating $(WORKTREE_ENV_FILE) with unique ports..."; \
+		bash scripts/init-worktree-env.sh $(WORKTREE_ENV_FILE); \
+	else \
+		echo "==> Using existing $(WORKTREE_ENV_FILE)"; \
+	fi
 	@$(MAKE) setup ENV_FILE=$(WORKTREE_ENV_FILE)
 
-start-worktree:
+start-worktree: ## Start this worktree using .env.worktree
 	@$(MAKE) start ENV_FILE=$(WORKTREE_ENV_FILE)
 
-stop-worktree:
+stop-worktree: ## Stop this worktree's backend and frontend processes
 	@$(MAKE) stop ENV_FILE=$(WORKTREE_ENV_FILE)
 
-check-worktree:
+check-worktree: ## Run the full verification pipeline for this worktree
 	@ENV_FILE=$(WORKTREE_ENV_FILE) bash scripts/check.sh
 
 # ---------- Individual commands ----------
+##@ Individual commands
 
-# Go server
-dev:
+dev: ## Bootstrap this checkout end-to-end: create env if needed, ensure DB, migrate, start services
+	@bash scripts/dev.sh
+
+server: ## Run only the Go server for the current checkout
 	$(REQUIRE_ENV)
 	@bash scripts/ensure-postgres.sh "$(ENV_FILE)"
 	cd server && go run ./cmd/server
 
-daemon:
-	@$(MAKE) multica MULTICA_ARGS="daemon"
+daemon: ## Restart the local agent daemon using the CLI's stored auth/session
+	@$(MAKE) multica MULTICA_ARGS="daemon restart --profile local"
 
-cli:
+cli: ## Run the multica CLI with ARGS or MULTICA_ARGS from source
 	@$(MAKE) multica MULTICA_ARGS="$(MULTICA_ARGS)"
 
-multica:
+multica: ## Run the multica CLI entrypoint directly from the Go source tree
 	cd server && go run ./cmd/multica $(MULTICA_ARGS)
 
 VERSION ?= $(shell git describe --tags --always --dirty 2>/dev/null || echo dev)
 COMMIT  ?= $(shell git rev-parse --short HEAD 2>/dev/null || echo unknown)
+DATE    ?= $(shell date -u '+%Y-%m-%dT%H:%M:%SZ')
 
-build:
-	cd server && go build -o bin/server ./cmd/server
-	cd server && go build -ldflags "-X main.version=$(VERSION) -X main.commit=$(COMMIT)" -o bin/multica ./cmd/multica
+build: ## Build the server, CLI, and migrate binaries into server/bin
+	cd server && go build -ldflags "-X main.version=$(VERSION) -X main.commit=$(COMMIT)" -o bin/server ./cmd/server
+	cd server && go build -ldflags "-X main.version=$(VERSION) -X main.commit=$(COMMIT) -X main.date=$(DATE)" -o bin/multica ./cmd/multica
+	cd server && go build -o bin/migrate ./cmd/migrate
 
-test:
+test: ## Run Go tests after ensuring the target DB exists and migrations are applied
 	$(REQUIRE_ENV)
 	@bash scripts/ensure-postgres.sh "$(ENV_FILE)"
+	cd server && go run ./cmd/migrate up
 	cd server && go test ./...
 
 # Database
-migrate-up:
+##@ Database
+
+migrate-up: ## Create the target DB if needed, then apply database migrations
 	$(REQUIRE_ENV)
 	@bash scripts/ensure-postgres.sh "$(ENV_FILE)"
 	cd server && go run ./cmd/migrate up
 
-migrate-down:
+migrate-down: ## Create the target DB if needed, then roll back database migrations
 	$(REQUIRE_ENV)
 	@bash scripts/ensure-postgres.sh "$(ENV_FILE)"
 	cd server && go run ./cmd/migrate down
 
-sqlc:
+sqlc: ## Regenerate sqlc code
 	cd server && sqlc generate
 
 # Cleanup
-clean:
+##@ Cleanup
+
+clean: ## Remove generated server binaries and temp files
 	rm -rf server/bin server/tmp

README.md

@@ -14,14 +14,13 @@
 
 **Your next 10 hires won't be human.**
 
-Open-source platform that turns coding agents into real teammates.<br/>
-Assign tasks, track progress, compound skills — manage your human + agent workforce in one place.
+The open-source managed agents platform.<br/>
+Turn coding agents into real teammates — assign tasks, track progress, compound skills.
 
 [![CI](https://github.com/multica-ai/multica/actions/workflows/ci.yml/badge.svg)](https://github.com/multica-ai/multica/actions/workflows/ci.yml)
-[![License](https://img.shields.io/badge/License-Apache_2.0-blue.svg)](https://opensource.org/licenses/Apache-2.0)
 [![GitHub stars](https://img.shields.io/github/stars/multica-ai/multica?style=flat)](https://github.com/multica-ai/multica/stargazers)
 
-[Website](https://multica.ai) · [Cloud](https://multica.ai/app) · [Self-Hosting](SELF_HOSTING.md) · [Contributing](CONTRIBUTING.md)
+[Website](https://multica.ai) · [Cloud](https://multica.ai/app) · [X](https://x.com/MulticaAI) · [Self-Hosting](SELF_HOSTING.md) · [Contributing](CONTRIBUTING.md)
 
 **English | [简体中文](README.zh-CN.md)**
 
@@ -31,71 +30,87 @@ Assign tasks, track progress, compound skills — manage your human + agent work
 
 Multica turns coding agents into real teammates. Assign issues to an agent like you'd assign to a colleague — they'll pick up the work, write code, report blockers, and update statuses autonomously.
 
-No more copy-pasting prompts. No more babysitting runs. Your agents show up on the board, participate in conversations, and compound reusable skills over time. Works with **Claude Code** and **Codex**.
+No more copy-pasting prompts. No more babysitting runs. Your agents show up on the board, participate in conversations, and compound reusable skills over time. Think of it as open-source infrastructure for managed agents — vendor-neutral, self-hosted, and designed for human + AI teams. Works with **Claude Code**, **Codex**, **GitHub Copilot CLI**, **OpenClaw**, **OpenCode**, **Hermes**, **Gemini**, **Pi**, **Cursor Agent**, **Kimi**, and **Kiro CLI**.
 
 <p align="center">
   <img src="docs/assets/hero-screenshot.png" alt="Multica board view" width="800">
 </p>
 
+## Why "Multica"?
+
+Multica — **Mul**tiplexed **I**nformation and **C**omputing **A**gent.
+
+The name is a nod to Multics, the pioneering operating system of the 1960s that introduced time-sharing — letting multiple users share a single machine as if each had it to themselves. Unix was born as a deliberate simplification of Multics: one user, one task, one elegant philosophy.
+
+We think the same inflection is happening again. For decades, software teams have been single-threaded — one engineer, one task, one context switch at a time. AI agents change that equation. Multica brings time-sharing back, but for an era where the "users" multiplexing the system are both humans and autonomous agents.
+
+In Multica, agents are first-class teammates. They get assigned issues, report progress, raise blockers, and ship code — just like their human colleagues. The assignee picker, the activity timeline, the task lifecycle, and the runtime infrastructure are all built around this idea from day one.
+
+Like Multics before it, the bet is on multiplexing: a small team shouldn't feel small. With the right system, two engineers and a fleet of agents can move like twenty.
+
 ## Features
 
+Multica manages the full agent lifecycle: from task assignment to execution monitoring to skill reuse.
+
 - **Agents as Teammates** — assign to an agent like you'd assign to a colleague. They have profiles, show up on the board, post comments, create issues, and report blockers proactively.
 - **Autonomous Execution** — set it and forget it. Full task lifecycle management (enqueue, claim, start, complete/fail) with real-time progress streaming via WebSocket.
 - **Reusable Skills** — every solution becomes a reusable skill for the whole team. Deployments, migrations, code reviews — skills compound your team's capabilities over time.
 - **Unified Runtimes** — one dashboard for all your compute. Local daemons and cloud runtimes, auto-detection of available CLIs, real-time monitoring.
 - **Multi-Workspace** — organize work across teams with workspace-level isolation. Each workspace has its own agents, issues, and settings.
 
+---
+
+## Quick Install
+
+### macOS / Linux (Homebrew - recommended)
+
+```bash
+brew install multica-ai/tap/multica
+```
+
+Use `brew upgrade multica-ai/tap/multica` to keep the CLI current.
+
+### macOS / Linux (install script)
+
+```bash
+curl -fsSL https://raw.githubusercontent.com/multica-ai/multica/main/scripts/install.sh | bash
+```
+
+Use this if Homebrew is not available. The script installs the Multica CLI on macOS and Linux by using Homebrew when it is on `PATH`, otherwise it downloads the binary directly.
+
+### Windows (PowerShell)
+
+```powershell
+irm https://raw.githubusercontent.com/multica-ai/multica/main/scripts/install.ps1 | iex
+```
+
+Then configure, authenticate, and start the daemon in one command:
+
+```bash
+multica setup          # Connect to Multica Cloud, log in, start daemon
+```
+
+> **Self-hosting?** Add `--with-server` to deploy a full Multica server on your machine:
+>
+> ```bash
+> curl -fsSL https://raw.githubusercontent.com/multica-ai/multica/main/scripts/install.sh | bash -s -- --with-server
+> multica setup self-host
+> ```
+>
+> This pulls the official Multica images from GHCR (latest stable by default). Requires Docker. See the [Self-Hosting Guide](SELF_HOSTING.md) for details.
+> If the selected GHCR tag has not been published yet, fall back to `make selfhost-build` from a checkout.
+
+---
+
 ## Getting Started
 
-### Multica Cloud
-
-The fastest way to get started — no setup required: **[multica.ai](https://multica.ai)**
-
-### Self-Host with Docker
+### 1. Set up and start the daemon
 
 ```bash
-git clone https://github.com/multica-ai/multica.git
-cd multica
-cp .env.example .env
-# Edit .env — at minimum, change JWT_SECRET
-
-docker compose up -d                              # Start PostgreSQL
-cd server && go run ./cmd/migrate up && cd ..     # Run migrations
-make start                                         # Start the app
+multica setup           # Configure, authenticate, and start the daemon
 ```
 
-See the [Self-Hosting Guide](SELF_HOSTING.md) for full instructions.
-
-## CLI
-
-The `multica` CLI connects your local machine to Multica — authenticate, manage workspaces, and run the agent daemon.
-
-```bash
-# Install
-brew tap multica-ai/tap
-brew install multica
-
-# Authenticate and start
-multica login
-multica daemon start
-```
-
-The daemon auto-detects available agent CLIs (`claude`, `codex`) on your PATH. When an agent is assigned a task, the daemon creates an isolated environment, runs the agent, and reports results back.
-
-See the [CLI and Daemon Guide](CLI_AND_DAEMON.md) for the full command reference, daemon configuration, and advanced usage.
-
-## Quickstart
-
-Once you have the CLI installed (or signed up for [Multica Cloud](https://multica.ai)), follow these steps to assign your first task to an agent:
-
-### 1. Log in and start the daemon
-
-```bash
-multica login           # Authenticate with your Multica account
-multica daemon start    # Start the local agent runtime
-```
-
-The daemon runs in the background and keeps your machine connected to Multica. It auto-detects agent CLIs (`claude`, `codex`) available on your PATH.
+The daemon runs in the background and auto-detects agent CLIs (`claude`, `codex`, `copilot`, `openclaw`, `opencode`, `hermes`, `gemini`, `pi`, `cursor-agent`, `kimi`, `kiro-cli`) on your PATH.
 
 ### 2. Verify your runtime
 
@@ -105,13 +120,47 @@ Open your workspace in the Multica web app. Navigate to **Settings → Runtimes*
 
 ### 3. Create an agent
 
-Go to **Settings → Agents** and click **New Agent**. Pick the runtime you just connected and choose a provider (Claude Code or Codex). Give your agent a name — this is how it will appear on the board, in comments, and in assignments.
+Go to **Settings → Agents** and click **New Agent**. Pick the runtime you just connected and choose a provider (Claude Code, Codex, GitHub Copilot CLI, OpenClaw, OpenCode, Hermes, Gemini, Pi, Cursor Agent, Kimi, or Kiro CLI). Give your agent a name — this is how it will appear on the board, in comments, and in assignments.
 
 ### 4. Assign your first task
 
 Create an issue from the board (or via `multica issue create`), then assign it to your new agent. The agent will automatically pick up the task, execute it on your runtime, and report progress — just like a human teammate.
 
-That's it! Your agent is now part of the team. 🎉
+---
+
+## Multica vs Paperclip
+
+| | Multica | Paperclip |
+|---|---------|-----------|
+| **Focus** | Team AI agent collaboration platform | Solo AI agent company simulator |
+| **User model** | Multi-user teams with roles & permissions | Single board operator |
+| **Agent interaction** | Issues + Chat conversations | Issues + Heartbeat |
+| **Deployment** | Cloud-first | Local-first |
+| **Management depth** | Lightweight (Issues / Projects / Labels) | Heavy governance (Org chart / Approvals / Budgets) |
+| **Extensibility** | Skills system | Skills + Plugin system |
+
+**TL;DR — Multica is built for teams that want to collaborate with AI agents on real projects together.**
+
+---
+
+## CLI
+
+The `multica` CLI connects your local machine to Multica — authenticate, manage workspaces, and run the agent daemon.
+
+| Command | Description |
+|---------|-------------|
+| `multica login` | Authenticate (opens browser) |
+| `multica daemon start` | Start the local agent runtime |
+| `multica daemon status` | Check daemon status |
+| `multica setup` | One-command setup for Multica Cloud (configure + login + start daemon) |
+| `multica setup self-host` | Same, but for self-hosted deployments |
+| `multica issue list` | List issues in your workspace |
+| `multica issue create` | Create a new issue |
+| `multica update` | Update to the latest version |
+
+See the [CLI and Daemon Guide](CLI_AND_DAEMON.md) for the full command reference.
+
+---
 
 ## Architecture
 
@@ -122,9 +171,10 @@ That's it! Your agent is now part of the team. 🎉
 └──────────────┘     └──────┬───────┘     └──────────────────┘
                             │
                      ┌──────┴───────┐
-                     │ Agent Daemon │  (runs on your machine)
-                     │ Claude/Codex │
-                     └──────────────┘
+                     │ Agent Daemon │  runs on your machine
+                     └──────────────┘  (Claude Code, Codex, GitHub Copilot CLI,
+                                        OpenCode, OpenClaw, Hermes, Gemini,
+                                        Pi, Cursor Agent, Kimi, Kiro CLI)
 ```
 
 | Layer | Stack |
@@ -132,7 +182,7 @@ That's it! Your agent is now part of the team. 🎉
 | Frontend | Next.js 16 (App Router) |
 | Backend | Go (Chi router, sqlc, gorilla/websocket) |
 | Database | PostgreSQL 17 with pgvector |
-| Agent Runtime | Local daemon executing Claude Code or Codex |
+| Agent Runtime | Local daemon executing Claude Code, Codex, GitHub Copilot CLI, OpenClaw, OpenCode, Hermes, Gemini, Pi, Cursor Agent, Kimi, or Kiro CLI |
 
 ## Development
 
@@ -141,14 +191,9 @@ For contributors working on the Multica codebase, see the [Contributing Guide](C
 **Prerequisites:** [Node.js](https://nodejs.org/) v20+, [pnpm](https://pnpm.io/) v10.28+, [Go](https://go.dev/) v1.26+, [Docker](https://www.docker.com/)
 
 ```bash
-pnpm install
-cp .env.example .env
-make setup
-make start
+make dev
 ```
 
+`make dev` auto-detects your environment (main checkout or worktree), creates the env file, installs dependencies, sets up the database, runs migrations, and starts all services.
+
 See [CONTRIBUTING.md](CONTRIBUTING.md) for the full development workflow, worktree support, testing, and troubleshooting.
-
-## License
-
-[Apache 2.0](LICENSE)

README.zh-CN.md

@@ -14,14 +14,13 @@
 
 **你的下一批员工，不是人类。**
 
-开源平台，将编码 Agent 变成真正的队友。<br/>
-分配任务、跟踪进度、积累技能——在一个地方管理你的人类 + Agent 团队。
+开源的 Managed Agents 平台。<br/>
+将编码 Agent 变成真正的队友——分配任务、跟踪进度、积累技能。
 
 [![CI](https://github.com/multica-ai/multica/actions/workflows/ci.yml/badge.svg)](https://github.com/multica-ai/multica/actions/workflows/ci.yml)
-[![License](https://img.shields.io/badge/License-Apache_2.0-blue.svg)](https://opensource.org/licenses/Apache-2.0)
 [![GitHub stars](https://img.shields.io/github/stars/multica-ai/multica?style=flat)](https://github.com/multica-ai/multica/stargazers)
 
-[官网](https://multica.ai) · [云服务](https://multica.ai/app) · [自部署指南](SELF_HOSTING.md) · [参与贡献](CONTRIBUTING.md)
+[官网](https://multica.ai) · [云服务](https://multica.ai/app) · [X](https://x.com/MulticaAI) · [自部署指南](SELF_HOSTING.md) · [参与贡献](CONTRIBUTING.md)
 
 **[English](README.md) | 简体中文**
 
@@ -31,71 +30,88 @@
 
 Multica 将编码 Agent 变成真正的队友。像分配给同事一样分配给 Agent——它们会自主接手工作、编写代码、报告阻塞问题、更新状态。
 
-不再需要复制粘贴 prompt，不再需要盯着运行过程。你的 Agent 出现在看板上、参与对话、随着时间积累可复用的技能。支持 **Claude Code** 和 **Codex**。
+不再需要复制粘贴 prompt，不再需要盯着运行过程。你的 Agent 出现在看板上、参与对话、随着时间积累可复用的技能。可以理解为开源的 Managed Agents 基础设施——厂商中立、可自部署、专为人类 + AI 团队设计。支持 **Claude Code**、**Codex**、**GitHub Copilot CLI**、**OpenClaw**、**OpenCode**、**Hermes**、**Gemini**、**Pi**、**Cursor Agent**、**Kimi** 和 **Kiro CLI**。
 
 <p align="center">
   <img src="docs/assets/hero-screenshot.png" alt="Multica 看板视图" width="800">
 </p>
 
+## 为什么叫 "Multica"？
+
+Multica——**Mul**tiplexed **I**nformation and **C**omputing **A**gent。
+
+这个名字是在向 20 世纪 60 年代具有开创意义的操作系统 Multics 致意。Multics 首创了分时系统，让多个用户能够共享同一台机器，同时又像各自独占它一样使用。Unix 则是在有意简化 Multics 的基础上诞生的，强调一个用户、一个任务、一种优雅的哲学。
+
+我们认为，类似的转折点正在再次出现。几十年来，软件团队一直处于一种单线程的工作模式，一个工程师处理一个任务，一次只专注于一个上下文。AI agents 改变了这个等式。Multica 将"分时"重新带回这个时代，只不过今天在系统中进行多路复用的"用户"，既包括人类，也包括自主代理。
+
+在 Multica 中，agents 是一级团队成员。它们会被分配 issue，汇报进展，提出阻塞，并交付代码，就像人类同事一样。任务分配、活动时间线、任务生命周期，以及运行时基础设施，Multica 从第一天起就是围绕这一理念构建的。
+
+和当年的 Multics 一样，这一判断建立在"多路复用"之上。一个小团队不该因为人数少就显得能力有限。有了合适的系统，两名工程师加上一组 agents，就能发挥出二十人团队的推进速度。
+
 ## 功能特性
 
+Multica 管理完整的 Agent 生命周期：从任务分配到执行监控再到技能复用。
+
 - **Agent 即队友** — 像分配给同事一样分配给 Agent。它们有个人档案、出现在看板上、发表评论、创建 Issue、主动报告阻塞问题。
 - **自主执行** — 设置后无需管理。完整的任务生命周期管理（排队、认领、执行、完成/失败），通过 WebSocket 实时推送进度。
 - **可复用技能** — 每个解决方案都成为全团队可复用的技能。部署、数据库迁移、代码审查——技能让团队能力随时间持续增长。
 - **统一运行时** — 一个控制台管理所有算力。本地 daemon 和云端运行时，自动检测可用 CLI，实时监控。
 - **多工作区** — 按团队组织工作，工作区级别隔离。每个工作区有独立的 Agent、Issue 和设置。
 
-## 快速开始
+---
 
-### Multica 云服务
+## 快速安装
 
-最快的上手方式，无需任何配置：**[multica.ai](https://multica.ai)**
-
-### Docker 自部署
+### macOS / Linux（推荐 Homebrew）
 
 ```bash
-git clone https://github.com/multica-ai/multica.git
-cd multica
-cp .env.example .env
-# 编辑 .env — 至少修改 JWT_SECRET
-
-docker compose up -d                              # 启动 PostgreSQL
-cd server && go run ./cmd/migrate up && cd ..     # 运行数据库迁移
-make start                                         # 启动应用
+brew install multica-ai/tap/multica
 ```
 
-完整部署文档请参阅 [自部署指南](SELF_HOSTING.md)。
+后续可用 `brew upgrade multica-ai/tap/multica` 更新 CLI。
 
-## CLI
-
-`multica` CLI 将你的本地机器连接到 Multica — 用于认证、管理工作区和运行 Agent daemon。
+### macOS / Linux（安装脚本）
 
 ```bash
-# 安装
-brew tap multica-ai/tap
-brew install multica
-
-# 认证并启动
-multica login
-multica daemon start
+curl -fsSL https://raw.githubusercontent.com/multica-ai/multica/main/scripts/install.sh | bash
 ```
 
-daemon 会自动检测 PATH 中可用的 Agent CLI（`claude`、`codex`）。当 Agent 被分配任务时，daemon 会创建隔离环境、运行 Agent、并将结果回传。
+如果没有 Homebrew，可以使用安装脚本。脚本会安装 Multica CLI：检测到 `brew` 时通过 Homebrew 安装，否则直接下载二进制。
 
-完整命令参考请参阅 [CLI 与 Daemon 指南](CLI_AND_DAEMON.md)。
+### Windows (PowerShell)
+
+```powershell
+irm https://raw.githubusercontent.com/multica-ai/multica/main/scripts/install.ps1 | iex
+```
+
+安装完成后，一条命令完成配置、认证和启动：
+
+```bash
+multica setup          # 连接 Multica Cloud，登录，启动 daemon
+```
+
+> **自部署？** 加上 `--with-server` 在本地部署完整的 Multica 服务：
+>
+> ```bash
+> curl -fsSL https://raw.githubusercontent.com/multica-ai/multica/main/scripts/install.sh | bash -s -- --with-server
+> multica setup self-host
+> ```
+>
+> 需要 Docker。详见 [自部署指南](SELF_HOSTING.md)。
+
+---
 
 ## 快速上手
 
 安装好 CLI（或注册 [Multica 云服务](https://multica.ai)）后，按以下步骤将第一个任务分配给 Agent：
 
-### 1. 登录并启动 daemon
+### 1. 配置并启动 daemon
 
 ```bash
-multica login           # 使用你的 Multica 账号认证
-multica daemon start    # 启动本地 Agent 运行时
+multica setup           # 配置、认证、启动 daemon（一条命令搞定）
 ```
 
-daemon 在后台运行，保持你的机器与 Multica 的连接。它会自动检测 PATH 中可用的 Agent CLI（`claude`、`codex`）。
+daemon 在后台运行，保持你的机器与 Multica 的连接。它会自动检测 PATH 中可用的 Agent CLI（`claude`、`codex`、`copilot`、`openclaw`、`opencode`、`hermes`、`gemini`、`pi`、`cursor-agent`、`kimi`、`kiro-cli`）。
 
 ### 2. 确认运行时已连接
 
@@ -105,7 +121,7 @@ daemon 在后台运行，保持你的机器与 Multica 的连接。它会自动
 
 ### 3. 创建 Agent
 
-进入 **设置 → Agents**，点击 **新建 Agent**。选择你刚连接的 Runtime，选择 Provider（Claude Code 或 Codex），并为 Agent 起个名字——它将以这个名字出现在看板、评论和任务分配中。
+进入 **设置 → Agents**，点击 **新建 Agent**。选择你刚连接的 Runtime，选择 Provider（Claude Code、Codex、GitHub Copilot CLI、OpenClaw、OpenCode、Hermes、Gemini、Pi、Cursor Agent、Kimi 或 Kiro CLI），并为 Agent 起个名字——它将以这个名字出现在看板、评论和任务分配中。
 
 ### 4. 分配你的第一个任务
 
@@ -113,6 +129,21 @@ daemon 在后台运行，保持你的机器与 Multica 的连接。它会自动
 
 大功告成！你的 Agent 现在是团队的一员了。 🎉
 
+---
+
+## Multica vs Paperclip
+
+| | Multica | Paperclip |
+|---|---------|-----------|
+| **定位** | 团队 AI Agent 协作平台 | 个人 AI Agent 公司模拟器 |
+| **用户模型** | 多人团队，角色权限 | 单人 Board Operator |
+| **Agent 交互** | Issue + Chat 对话 | Issue + Heartbeat |
+| **部署** | 云端优先 | 本地优先 |
+| **管理深度** | 轻量（Issue / Project / Labels） | 重度（组织架构 / 审批 / 预算） |
+| **扩展** | Skills 系统 | Skills + 插件系统 |
+
+**简单来说：Multica 专为团队协作打造，让团队和 AI Agent 一起高效完成项目。**
+
 ## 架构
 
 ```
@@ -122,9 +153,10 @@ daemon 在后台运行，保持你的机器与 Multica 的连接。它会自动
 └──────────────┘     └──────┬───────┘     └──────────────────┘
                             │
                      ┌──────┴───────┐
-                     │ Agent Daemon │  （运行在你的机器上）
-                     │ Claude/Codex │
-                     └──────────────┘
+                     │ Agent Daemon │  运行在你的机器上
+                     └──────────────┘  （Claude Code、Codex、GitHub Copilot CLI、
+                                        OpenCode、OpenClaw、Hermes、Gemini、
+                                        Pi、Cursor Agent、Kimi、Kiro CLI）
 ```
 
 | 层级 | 技术栈 |
@@ -132,7 +164,7 @@ daemon 在后台运行，保持你的机器与 Multica 的连接。它会自动
 | 前端 | Next.js 16 (App Router) |
 | 后端 | Go (Chi router, sqlc, gorilla/websocket) |
 | 数据库 | PostgreSQL 17 with pgvector |
-| Agent 运行时 | 本地 daemon 执行 Claude Code 或 Codex |
+| Agent 运行时 | 本地 daemon 执行 Claude Code、Codex、GitHub Copilot CLI、OpenClaw、OpenCode、Hermes、Gemini、Pi、Cursor Agent、Kimi 或 Kiro CLI |
 
 ## 开发
 

SELF_HOSTING.md

@@ -1,10 +1,8 @@
 # Self-Hosting Guide
 
-This guide walks you through deploying Multica on your own infrastructure.
+Deploy Multica on your own infrastructure in minutes.
 
-## Architecture Overview
-
-Multica has three components:
+## Architecture
 
 | Component | Description | Technology |
 |-----------|-------------|------------|
@@ -12,16 +10,176 @@ Multica has three components:
 | **Frontend** | Web application | Next.js 16 |
 | **Database** | Primary data store | PostgreSQL 17 with pgvector |
 
-Additionally, each user who wants to run AI agents locally installs the **`multica` CLI** and runs the **agent daemon** on their own machine.
+Each user who runs AI agents locally also installs the **`multica` CLI** and runs the **agent daemon** on their own machine.
 
-## Prerequisites
+## Quick Install (Recommended)
 
-- Docker and Docker Compose (recommended), or:
-  - Go 1.26+ (to build from source)
-  - Node.js 20+ and pnpm 10.28+ (to build the frontend)
-  - PostgreSQL 17 with the pgvector extension
+Two commands to set up everything — server, CLI, and configuration:
 
-## Quick Start (Docker Compose)
+```bash
+# 1. Install CLI + provision the self-host server
+curl -fsSL https://raw.githubusercontent.com/multica-ai/multica/main/scripts/install.sh | bash -s -- --with-server
+
+# 2. Configure CLI, authenticate, and start the daemon
+multica setup self-host
+```
+
+This installs the `multica` CLI, checks out the latest self-host assets, pulls the official Multica images from GHCR, and configures everything for localhost.
+
+Open http://localhost:3000. To log in, configure `RESEND_API_KEY` in `.env` for email-based codes (recommended), or leave Resend unset and copy the generated code from the backend logs. See [Step 2 — Log In](#step-2--log-in) for details.
+
+> **Prerequisites:** Docker and Docker Compose must be installed. The script checks for this and provides install links if missing.
+>
+> **CLI only?** If the self-host server is already running and you only need the CLI on a macOS/Linux machine, install it with Homebrew:
+>
+> ```bash
+> brew install multica-ai/tap/multica
+> ```
+
+---
+
+## Step-by-Step Setup (Alternative)
+
+If you prefer to run each step manually:
+
+### Step 1 — Start the Server
+
+**Prerequisites:** Docker and Docker Compose.
+
+```bash
+git clone https://github.com/multica-ai/multica.git
+cd multica
+make selfhost
+```
+
+`make selfhost` automatically creates `.env` from the example, generates a random `JWT_SECRET`, and starts all services via Docker Compose.
+
+By default it pulls the latest stable release images from GHCR. To build the backend/web from your current checkout instead, run `make selfhost-build`.
+If the selected GHCR tag has not been published yet, `make selfhost` now tells you to fall back to `make selfhost-build`.
+`make selfhost-build` uses local `multica-backend:dev` / `multica-web:dev` tags, so it does not overwrite the pulled `:latest` images.
+
+Once ready:
+
+- **Frontend:** http://localhost:3000
+- **Backend API:** http://localhost:8080
+
+> **Note:** If you prefer to run the Docker Compose steps manually, see [Manual Docker Compose Setup](#manual-docker-compose-setup) below.
+
+### Step 2 — Log In
+
+Open http://localhost:3000 in your browser. The Docker self-host stack defaults to `APP_ENV=production` (set in `docker-compose.selfhost.yml`), and there is no fixed verification code by default. Pick one of the following to log in:
+
+- **Recommended (production):** configure `RESEND_API_KEY` in `.env`, then restart the backend. Real verification codes will be sent to the email address you enter. See [Advanced Configuration → Email](SELF_HOSTING_ADVANCED.md#email-required-for-authentication).
+- **Without email configured:** the verification code is generated server-side and printed to the backend container logs (look for `[DEV] Verification code for ...:`). Useful for one-off testing on a single machine.
+- **Deterministic local/private testing:** set `APP_ENV=development` and `MULTICA_DEV_VERIFICATION_CODE=888888` in `.env`, then restart the backend. This fixed code is ignored when `APP_ENV=production`.
+
+Changes to `ALLOW_SIGNUP` and `GOOGLE_CLIENT_ID` also take effect after restarting the backend / compose stack. The web UI reads both from `/api/config` at runtime, so no web rebuild is needed.
+
+> **Warning:** do **not** set `MULTICA_DEV_VERIFICATION_CODE` on a publicly reachable instance — anyone who knows an email address can then log in with that fixed code.
+
+### Step 3 — Install CLI & Start Daemon
+
+The daemon runs on your local machine (not inside Docker). It detects installed AI agent CLIs, registers them with the server, and executes tasks when agents are assigned work.
+
+Each team member who wants to run AI agents locally needs to:
+
+### a) Install the CLI and an AI agent
+
+```bash
+brew install multica-ai/tap/multica
+```
+
+You also need at least one AI agent CLI installed:
+- [Claude Code](https://docs.anthropic.com/en/docs/claude-code) (`claude` on PATH)
+- [Codex](https://github.com/openai/codex) (`codex` on PATH)
+- [GitHub Copilot CLI](https://docs.github.com/en/copilot) (`copilot` on PATH)
+- [OpenClaw](https://github.com/openclaw/openclaw) (`openclaw` on PATH)
+- [OpenCode](https://github.com/anomalyco/opencode) (`opencode` on PATH)
+- [Hermes](https://github.com/NousResearch/hermes) (`hermes` on PATH)
+- Gemini (`gemini` on PATH)
+- [Pi](https://pi.dev/) (`pi` on PATH)
+- [Cursor Agent](https://cursor.com/) (`cursor-agent` on PATH)
+- Kimi (`kimi` on PATH)
+- Kiro CLI (`kiro-cli` on PATH)
+
+### b) One-command setup
+
+```bash
+multica setup self-host
+```
+
+This automatically:
+1. Configures the CLI to connect to `localhost` (ports 8080/3000)
+2. Opens your browser for authentication
+3. Discovers your workspaces
+4. Starts the daemon in the background
+
+For on-premise deployments with custom domains:
+
+```bash
+multica setup self-host --server-url https://api.example.com --app-url https://app.example.com
+```
+
+To verify the daemon is running:
+
+```bash
+multica daemon status
+```
+
+> **Alternative:** If you prefer manual steps, see [Manual CLI Configuration](#manual-cli-configuration) below.
+
+### Step 4 — Verify & Start Using
+
+1. Open your workspace in the web app at http://localhost:3000
+2. Navigate to **Settings → Runtimes** — you should see your machine listed
+3. Go to **Settings → Agents** and create a new agent
+4. Create an issue and assign it to your agent — it will pick up the task automatically
+
+## Stopping Services
+
+If you installed via the install script:
+
+```bash
+curl -fsSL https://raw.githubusercontent.com/multica-ai/multica/main/scripts/install.sh | bash -s -- --stop
+```
+
+If you cloned the repo manually:
+
+```bash
+# Stop the Docker Compose services (backend, frontend, database)
+make selfhost-stop
+
+# Stop the local daemon
+multica daemon stop
+```
+
+## Switching to Multica Cloud
+
+If you've been self-hosting and want to switch your CLI to [Multica Cloud](https://multica.ai):
+
+```bash
+multica setup
+```
+
+This reconfigures the CLI for multica.ai, re-authenticates, and restarts the daemon. You will be prompted before overwriting the existing configuration.
+
+> Your local Docker services are unaffected. Stop them separately if you no longer need them.
+
+## Upgrading
+
+```bash
+docker compose -f docker-compose.selfhost.yml pull
+docker compose -f docker-compose.selfhost.yml up -d
+```
+
+Pin `MULTICA_IMAGE_TAG` in `.env` to an exact version like `v0.2.4` if you want to stay on a specific release. Migrations run automatically on backend startup.
+If the selected GHCR tag has not been published yet, fall back to `make selfhost-build` or `docker compose -f docker-compose.selfhost.yml -f docker-compose.selfhost.build.yml up -d --build`.
+
+---
+
+## Manual Docker Compose Setup
+
+If you prefer running Docker Compose steps manually instead of `make selfhost`:
 
 ```bash
 git clone https://github.com/multica-ai/multica.git
@@ -29,250 +187,44 @@ cd multica
 cp .env.example .env
 ```
 
-Edit `.env` with your production values (see [Configuration](#configuration) below), then:
+Edit `.env` — at minimum, change `JWT_SECRET`:
 
 ```bash
-# Start PostgreSQL
-docker compose up -d
-
-# Build the backend
-make build
-
-# Run database migrations
-DATABASE_URL="your-database-url" ./server/bin/migrate up
-
-# Start the backend server
-DATABASE_URL="your-database-url" PORT=8080 ./server/bin/server
+JWT_SECRET=$(openssl rand -hex 32)
 ```
 
-For the frontend:
+Then start everything:
 
 ```bash
-pnpm install
-pnpm build
-
-# Start the frontend (production mode)
-cd apps/web
-REMOTE_API_URL=http://localhost:8080 pnpm start
+docker compose -f docker-compose.selfhost.yml pull
+docker compose -f docker-compose.selfhost.yml up -d
 ```
 
-## Configuration
+## Manual CLI Configuration
 
-All configuration is done via environment variables. Copy `.env.example` as a starting point.
-
-### Required Variables
-
-| Variable | Description | Example |
-|----------|-------------|---------|
-| `DATABASE_URL` | PostgreSQL connection string | `postgres://multica:multica@localhost:5432/multica?sslmode=disable` |
-| `JWT_SECRET` | **Must change from default.** Secret key for signing JWT tokens. Use a long random string. | `openssl rand -hex 32` |
-| `FRONTEND_ORIGIN` | URL where the frontend is served (used for CORS) | `https://app.example.com` |
-
-### Email (Required for Authentication)
-
-Multica uses email-based magic link authentication via [Resend](https://resend.com).
-
-| Variable | Description |
-|----------|-------------|
-| `RESEND_API_KEY` | Your Resend API key |
-| `RESEND_FROM_EMAIL` | Sender email address (default: `noreply@multica.ai`) |
-
-### Google OAuth (Optional)
-
-| Variable | Description |
-|----------|-------------|
-| `GOOGLE_CLIENT_ID` | Google OAuth client ID |
-| `GOOGLE_CLIENT_SECRET` | Google OAuth client secret |
-| `GOOGLE_REDIRECT_URI` | OAuth callback URL (e.g. `https://app.example.com/auth/callback`) |
-
-### File Storage (Optional)
-
-For file uploads and attachments, configure S3 and CloudFront:
-
-| Variable | Description |
-|----------|-------------|
-| `S3_BUCKET` | S3 bucket name |
-| `S3_REGION` | AWS region (default: `us-west-2`) |
-| `CLOUDFRONT_DOMAIN` | CloudFront distribution domain |
-| `CLOUDFRONT_KEY_PAIR_ID` | CloudFront key pair ID for signed URLs |
-| `CLOUDFRONT_PRIVATE_KEY` | CloudFront private key (PEM format) |
-| `COOKIE_DOMAIN` | Domain for CloudFront auth cookies |
-
-### Server
-
-| Variable | Default | Description |
-|----------|---------|-------------|
-| `PORT` | `8080` | Backend server port |
-| `FRONTEND_PORT` | `3000` | Frontend port |
-| `CORS_ALLOWED_ORIGINS` | Value of `FRONTEND_ORIGIN` | Comma-separated list of allowed origins |
-| `LOG_LEVEL` | `info` | Log level: `debug`, `info`, `warn`, `error` |
-
-### CLI / Daemon
-
-These are configured on each user's machine, not on the server:
-
-| Variable | Default | Description |
-|----------|---------|-------------|
-| `MULTICA_SERVER_URL` | `ws://localhost:8080/ws` | WebSocket URL for daemon → server connection |
-| `MULTICA_APP_URL` | `http://localhost:3000` | Frontend URL for CLI login flow |
-| `MULTICA_DAEMON_POLL_INTERVAL` | `3s` | How often the daemon polls for tasks |
-| `MULTICA_DAEMON_HEARTBEAT_INTERVAL` | `15s` | Heartbeat frequency |
-
-## Database Setup
-
-Multica requires PostgreSQL 17 with the pgvector extension.
-
-### Using the Included Docker Compose
+If you prefer configuring the CLI step by step instead of `multica setup`:
 
 ```bash
-docker compose up -d postgres
+# Point CLI to your local server
+multica config set server_url http://localhost:8080
+multica config set app_url http://localhost:3000
+
+# Login (opens browser)
+multica login
+
+# Start the daemon
+multica daemon start
 ```
 
-This starts a `pgvector/pgvector:pg17` container on port 5432 with default credentials (`multica`/`multica`).
-
-### Using Your Own PostgreSQL
-
-Ensure the pgvector extension is available:
-
-```sql
-CREATE EXTENSION IF NOT EXISTS vector;
-```
-
-### Running Migrations
-
-Migrations must be run before starting the server:
+For production deployments with TLS:
 
 ```bash
-# Using the built binary
-./server/bin/migrate up
-
-# Or from source
-cd server && go run ./cmd/migrate up
+multica config set app_url https://app.example.com
+multica config set server_url https://api.example.com
+multica login
+multica daemon start
 ```
 
-## Reverse Proxy
+## Advanced Configuration
 
-In production, put a reverse proxy in front of both the backend and frontend to handle TLS and routing.
-
-### Caddy (Recommended)
-
-```
-app.example.com {
-    reverse_proxy localhost:3000
-}
-
-api.example.com {
-    reverse_proxy localhost:8080
-}
-```
-
-### Nginx
-
-```nginx
-# Frontend
-server {
-    listen 443 ssl;
-    server_name app.example.com;
-
-    ssl_certificate     /path/to/cert.pem;
-    ssl_certificate_key /path/to/key.pem;
-
-    location / {
-        proxy_pass http://localhost:3000;
-        proxy_set_header Host $host;
-        proxy_set_header X-Real-IP $remote_addr;
-        proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
-        proxy_set_header X-Forwarded-Proto $scheme;
-    }
-}
-
-# Backend API
-server {
-    listen 443 ssl;
-    server_name api.example.com;
-
-    ssl_certificate     /path/to/cert.pem;
-    ssl_certificate_key /path/to/key.pem;
-
-    location / {
-        proxy_pass http://localhost:8080;
-        proxy_set_header Host $host;
-        proxy_set_header X-Real-IP $remote_addr;
-        proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
-        proxy_set_header X-Forwarded-Proto $scheme;
-    }
-
-    # WebSocket support
-    location /ws {
-        proxy_pass http://localhost:8080;
-        proxy_http_version 1.1;
-        proxy_set_header Upgrade $http_upgrade;
-        proxy_set_header Connection "upgrade";
-        proxy_set_header Host $host;
-        proxy_read_timeout 86400;
-    }
-}
-```
-
-When using separate domains for frontend and backend, set these environment variables accordingly:
-
-```bash
-# Backend
-FRONTEND_ORIGIN=https://app.example.com
-CORS_ALLOWED_ORIGINS=https://app.example.com
-
-# Frontend
-REMOTE_API_URL=https://api.example.com
-NEXT_PUBLIC_API_URL=https://api.example.com
-NEXT_PUBLIC_WS_URL=wss://api.example.com/ws
-```
-
-## Health Check
-
-The backend exposes a health check endpoint:
-
-```
-GET /health
-→ {"status":"ok"}
-```
-
-Use this for load balancer health checks or monitoring.
-
-## Setting Up the Agent Daemon
-
-Each team member who wants to run AI agents locally needs to:
-
-1. **Install the CLI**
-
-   ```bash
-   brew tap multica-ai/tap
-   brew install multica-cli
-   ```
-
-2. **Install an AI agent CLI** — at least one of:
-   - [Claude Code](https://docs.anthropic.com/en/docs/claude-code) (`claude` on PATH)
-   - [Codex](https://github.com/openai/codex) (`codex` on PATH)
-
-3. **Authenticate and start**
-
-   ```bash
-   # Point CLI to your server
-   export MULTICA_APP_URL=https://app.example.com
-   export MULTICA_SERVER_URL=wss://api.example.com/ws
-
-   # Login (opens browser)
-   multica login
-
-   # Start the daemon
-   multica daemon start
-   ```
-
-The daemon auto-detects installed agent CLIs and registers itself with the server. When an agent is assigned a task in Multica, the daemon picks it up, creates an isolated workspace, runs the agent, and reports results back.
-
-## Upgrading
-
-1. Pull the latest code or image
-2. Run migrations: `./server/bin/migrate up`
-3. Restart the backend and frontend
-
-Migrations are forward-only and safe to run on a live database. They are idempotent — running them multiple times has no effect.
+For environment variables, manual setup (without Docker), reverse proxy configuration, database setup, and more, see the [Advanced Configuration Guide](SELF_HOSTING_ADVANCED.md).

SELF_HOSTING_ADVANCED.md

@@ -0,0 +1,346 @@
+# Self-Hosting — Advanced Configuration
+
+This document covers advanced configuration for self-hosted Multica deployments. For the quick start guide, see [SELF_HOSTING.md](SELF_HOSTING.md).
+
+## Configuration
+
+All configuration is done via environment variables. Copy `.env.example` as a starting point.
+
+### Required Variables
+
+| Variable | Description | Example |
+|----------|-------------|---------|
+| `DATABASE_URL` | PostgreSQL connection string | `postgres://multica:multica@localhost:5432/multica?sslmode=disable` |
+| `JWT_SECRET` | **Must change from default.** Secret key for signing JWT tokens. Use a long random string. | `openssl rand -hex 32` |
+| `FRONTEND_ORIGIN` | URL where the frontend is served (used for CORS) | `https://app.example.com` |
+
+### Database Pool Tuning (Optional)
+
+These have sensible defaults and only need to be set when tuning a large or constrained deployment. Precedence (highest first): env var → `pool_*` query params on `DATABASE_URL` → built-in default.
+
+| Variable | Description | Default |
+|----------|-------------|---------|
+| `DATABASE_MAX_CONNS` | pgxpool max connections per pod. `pod_count × DATABASE_MAX_CONNS` should stay well below the Postgres `max_connections` ceiling. With a connection pooler (PgBouncer / RDS Proxy / Supavisor) in front, this can be raised significantly. | `25` |
+| `DATABASE_MIN_CONNS` | pgxpool warm baseline connections per pod. Auto-clamped to `DATABASE_MAX_CONNS`. | `5` |
+
+### Email (Required for Authentication)
+
+Multica uses email-based magic link authentication via [Resend](https://resend.com).
+
+| Variable | Description |
+|----------|-------------|
+| `RESEND_API_KEY` | Your Resend API key |
+| `RESEND_FROM_EMAIL` | Sender email address (default: `noreply@multica.ai`) |
+
+> **Note:** If Resend is not configured, generated verification codes are printed to backend logs. A fixed local testing code is disabled by default; to opt in on a private test instance, set `APP_ENV=development` and `MULTICA_DEV_VERIFICATION_CODE` to a 6-digit value. It is ignored when `APP_ENV=production`.
+
+### Google OAuth (Optional)
+
+| Variable | Description |
+|----------|-------------|
+| `GOOGLE_CLIENT_ID` | Google OAuth client ID |
+| `GOOGLE_CLIENT_SECRET` | Google OAuth client secret |
+| `GOOGLE_REDIRECT_URI` | OAuth callback URL (e.g. `https://app.example.com/auth/callback`) |
+
+Changes take effect after restarting the backend / compose stack. The web UI reads `GOOGLE_CLIENT_ID` from `/api/config` at runtime, so no web rebuild is needed.
+
+### Signup Controls (Optional)
+
+| Variable | Description |
+|----------|-------------|
+| `ALLOW_SIGNUP` | Set to `false` to disable new user signups on a private instance |
+| `ALLOWED_EMAIL_DOMAINS` | Optional comma-separated allowlist of email domains |
+| `ALLOWED_EMAILS` | Optional comma-separated allowlist of exact email addresses |
+
+Changes take effect after restarting the backend / compose stack. The web UI reads `ALLOW_SIGNUP` from `/api/config` at runtime, so no web rebuild is needed.
+
+### File Storage (Optional)
+
+For file uploads and attachments, configure S3 and (optionally) CloudFront:
+
+| Variable | Description |
+|----------|-------------|
+| `S3_BUCKET` | Bucket name only (e.g. `my-bucket`). Do **not** include the `.s3.<region>.amazonaws.com` suffix — the server constructs the public URL from `S3_BUCKET` + `S3_REGION` |
+| `S3_REGION` | AWS region (default: `us-west-2`). Must match the bucket's actual region — used for both SDK signing and public URLs |
+| `AWS_ACCESS_KEY_ID` / `AWS_SECRET_ACCESS_KEY` | Static credentials. When both are unset, the AWS SDK default credential chain is used |
+| `AWS_ENDPOINT_URL` | Custom S3-compatible endpoint (e.g. MinIO, R2, B2). Setting this switches the public URL to path-style |
+| `CLOUDFRONT_DOMAIN` | CloudFront distribution domain — when set, public URLs use this host instead of the S3 host |
+| `CLOUDFRONT_KEY_PAIR_ID` | CloudFront key pair ID for signed URLs |
+| `CLOUDFRONT_PRIVATE_KEY` | CloudFront private key (PEM format) |
+
+### Cookies
+
+| Variable | Description |
+|----------|-------------|
+| `COOKIE_DOMAIN` | Optional `Domain` attribute for session + CloudFront cookies. **Leave empty** for single-host deployments (localhost, LAN IP, or a single hostname). Only set it when the frontend and backend sit on different subdomains of one registered domain (e.g. `.example.com`). **Do not use an IP literal** — RFC 6265 forbids IP addresses in the cookie `Domain` attribute and browsers will drop such `Set-Cookie` headers. |
+
+The `Secure` flag on session cookies is derived automatically from the scheme of `FRONTEND_ORIGIN`: HTTPS origins get `Secure` cookies; plain-HTTP origins (LAN / private-network self-host) get non-secure cookies so the browser can actually store them.
+
+### Server
+
+| Variable | Default | Description |
+|----------|---------|-------------|
+| `PORT` | `8080` | Backend server port |
+| `METRICS_ADDR` | empty | Optional Prometheus metrics listener, for example `127.0.0.1:9090` |
+| `FRONTEND_PORT` | `3000` | Frontend port |
+| `CORS_ALLOWED_ORIGINS` | Value of `FRONTEND_ORIGIN` | Comma-separated list of allowed origins |
+| `LOG_LEVEL` | `info` | Log level: `debug`, `info`, `warn`, `error` |
+
+### CLI / Daemon
+
+These are configured on each user's machine, not on the server:
+
+| Variable | Default | Description |
+|----------|---------|-------------|
+| `MULTICA_SERVER_URL` | `ws://localhost:8080/ws` | WebSocket URL for daemon → server connection |
+| `MULTICA_APP_URL` | `http://localhost:3000` | Frontend URL for CLI login flow |
+| `MULTICA_DAEMON_POLL_INTERVAL` | `3s` | How often the daemon polls for tasks |
+| `MULTICA_DAEMON_HEARTBEAT_INTERVAL` | `15s` | Heartbeat frequency |
+
+Agent-specific overrides:
+
+| Variable | Description |
+|----------|-------------|
+| `MULTICA_CLAUDE_PATH` | Custom path to the `claude` binary |
+| `MULTICA_CLAUDE_MODEL` | Override the Claude model used |
+| `MULTICA_CODEX_PATH` | Custom path to the `codex` binary |
+| `MULTICA_CODEX_MODEL` | Override the Codex model used |
+| `MULTICA_COPILOT_PATH` | Custom path to the `copilot` (GitHub Copilot CLI) binary |
+| `MULTICA_COPILOT_MODEL` | Override the Copilot model used (note: GitHub Copilot routes models through your account entitlement, so this may not be honoured) |
+| `MULTICA_OPENCODE_PATH` | Custom path to the `opencode` binary |
+| `MULTICA_OPENCODE_MODEL` | Override the OpenCode model used |
+| `MULTICA_OPENCLAW_PATH` | Custom path to the `openclaw` binary |
+| `MULTICA_OPENCLAW_MODEL` | Override the OpenClaw model used |
+| `MULTICA_HERMES_PATH` | Custom path to the `hermes` binary |
+| `MULTICA_HERMES_MODEL` | Override the Hermes model used |
+| `MULTICA_GEMINI_PATH` | Custom path to the `gemini` binary |
+| `MULTICA_GEMINI_MODEL` | Override the Gemini model used |
+| `MULTICA_PI_PATH` | Custom path to the `pi` binary |
+| `MULTICA_PI_MODEL` | Override the Pi model used |
+| `MULTICA_CURSOR_PATH` | Custom path to the `cursor-agent` binary |
+| `MULTICA_CURSOR_MODEL` | Override the Cursor Agent model used |
+
+## Database Setup
+
+Multica requires PostgreSQL 17 with the pgvector extension.
+
+### Using Docker Compose (Recommended)
+
+The `docker-compose.selfhost.yml` includes PostgreSQL. No separate setup needed.
+
+### Using Your Own PostgreSQL
+
+If you prefer to use an existing PostgreSQL instance, ensure the pgvector extension is available:
+
+```sql
+CREATE EXTENSION IF NOT EXISTS vector;
+```
+
+Set `DATABASE_URL` in your `.env` and remove the `postgres` service from the compose file.
+
+### Running Migrations Manually
+
+The Docker Compose setup runs migrations automatically. If you need to run them manually:
+
+```bash
+# Using the built binary
+./server/bin/migrate up
+
+# Or from source
+cd server && go run ./cmd/migrate up
+```
+
+## Manual Setup (Without Docker Compose)
+
+If you prefer to build and run services manually:
+
+**Prerequisites:** Go 1.26+, Node.js 20+, pnpm 10.28+, PostgreSQL 17 with pgvector.
+
+```bash
+# Start your PostgreSQL (or use: docker compose up -d postgres)
+
+# Build the backend
+make build
+
+# Run database migrations
+DATABASE_URL="your-database-url" ./server/bin/migrate up
+
+# Start the backend server
+DATABASE_URL="your-database-url" PORT=8080 JWT_SECRET="your-secret" ./server/bin/server
+```
+
+For the frontend:
+
+```bash
+pnpm install
+pnpm build
+
+# Start the frontend (production mode)
+cd apps/web
+REMOTE_API_URL=http://localhost:8080 pnpm start
+```
+
+## Reverse Proxy
+
+In production, put a reverse proxy in front of both the backend and frontend to handle TLS and routing.
+
+### Caddy (Recommended)
+
+```
+app.example.com {
+    reverse_proxy localhost:3000
+}
+
+api.example.com {
+    reverse_proxy localhost:8080
+}
+```
+
+### Nginx
+
+```nginx
+# Frontend
+server {
+    listen 443 ssl;
+    server_name app.example.com;
+
+    ssl_certificate     /path/to/cert.pem;
+    ssl_certificate_key /path/to/key.pem;
+
+    location / {
+        proxy_pass http://localhost:3000;
+        proxy_set_header Host $host;
+        proxy_set_header X-Real-IP $remote_addr;
+        proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
+        proxy_set_header X-Forwarded-Proto $scheme;
+    }
+}
+
+# Backend API
+server {
+    listen 443 ssl;
+    server_name api.example.com;
+
+    ssl_certificate     /path/to/cert.pem;
+    ssl_certificate_key /path/to/key.pem;
+
+    location / {
+        proxy_pass http://localhost:8080;
+        proxy_set_header Host $host;
+        proxy_set_header X-Real-IP $remote_addr;
+        proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
+        proxy_set_header X-Forwarded-Proto $scheme;
+    }
+
+    # WebSocket support
+    location /ws {
+        proxy_pass http://localhost:8080;
+        proxy_http_version 1.1;
+        proxy_set_header Upgrade $http_upgrade;
+        proxy_set_header Connection "upgrade";
+        proxy_set_header Host $host;
+        proxy_read_timeout 86400;
+    }
+}
+```
+
+When using separate domains for frontend and backend, set these environment variables accordingly:
+
+```bash
+# Backend
+FRONTEND_ORIGIN=https://app.example.com
+CORS_ALLOWED_ORIGINS=https://app.example.com
+
+# Frontend (only if you are building the web image from source via docker-compose.selfhost.build.yml)
+REMOTE_API_URL=https://api.example.com
+NEXT_PUBLIC_API_URL=https://api.example.com
+NEXT_PUBLIC_WS_URL=wss://api.example.com/ws
+```
+
+## LAN / Non-localhost Access
+
+By default, Multica works on `localhost`. If you access it from another machine on the LAN (e.g. `http://192.168.1.100:3000`), you need to tell the backend to accept that origin:
+
+```bash
+# .env — replace with your server's LAN IP
+FRONTEND_ORIGIN=http://192.168.1.100:3000
+CORS_ALLOWED_ORIGINS=http://192.168.1.100:3000
+```
+
+Then restart the stack:
+
+```bash
+docker compose -f docker-compose.selfhost.yml up -d
+```
+
+### WebSocket for LAN / Non-localhost Access
+
+HTTP requests (issues, comments, uploads) work on LAN out of the box — Next.js rewrites proxy `/api`, `/auth`, and `/uploads` to the backend. **WebSockets do not**: Next.js rewrites only forward HTTP requests, not the `Upgrade` handshake a WebSocket needs. If you open the app on `http://<lan-ip>:3000`, real-time features (chat streaming, live issue updates, notifications) will fail to connect until you do one of the following:
+
+1. **Put a reverse proxy in front of the stack (recommended).** Nginx or Caddy terminates the WebSocket upgrade and forwards it to the backend on port 8080. See the [Reverse Proxy](#reverse-proxy) section above — the Nginx example already includes a `location /ws { ... }` block with the correct `Upgrade` / `Connection` headers. Once a proxy is in place the browser connects directly through it, so no frontend rebuild is needed.
+
+2. **Bake a WebSocket URL into the web image.** If you are not running a reverse proxy, rebuild the web image with `NEXT_PUBLIC_WS_URL` pointing straight at the backend (port 8080 must be reachable from the browser):
+
+   ```bash
+   # In .env
+   NEXT_PUBLIC_WS_URL=ws://<lan-ip>:8080/ws
+
+   # Rebuild the web image so the build-time value is baked in
+   docker compose -f docker-compose.selfhost.yml -f docker-compose.selfhost.build.yml up -d --build
+   ```
+
+   `NEXT_PUBLIC_WS_URL` is a build-time variable (see `Dockerfile.web`), so setting it only in `environment:` on the pre-built image has no effect — you must use the `selfhost.build.yml` override that rebuilds the image.
+
+> **Note:** If you need to hard-code a different public API / WebSocket endpoint into the web image for any other reason, use the same source-build override: `docker compose -f docker-compose.selfhost.yml -f docker-compose.selfhost.build.yml up -d --build`.
+
+## Health Check
+
+The backend exposes public health endpoints:
+
+```text
+GET /health
+→ {"status":"ok"}
+
+GET /readyz
+→ {"status":"ok","checks":{"db":"ok","migrations":"ok"}}
+
+GET /healthz
+→ same response as /readyz
+```
+
+Use `/health` for basic liveness / reachability checks. Use `/readyz` for
+dependency-aware readiness probes and external monitoring that should fail when
+the database is unavailable or migrations are not fully applied. `/healthz` is
+kept as an alias for operator familiarity.
+
+## Prometheus Metrics
+
+The backend can expose Prometheus metrics on a separate management listener:
+
+```bash
+METRICS_ADDR=127.0.0.1:9090 ./server/bin/server
+curl http://127.0.0.1:9090/metrics
+```
+
+`METRICS_ADDR` is empty by default, so no metrics listener is started. The
+public API port does not serve `/metrics`; keep it that way for internet-facing
+deployments. HTTP request metrics start accumulating only after the metrics
+listener is enabled. Metrics can reveal internal routes, traffic volume,
+dependency state, and runtime health.
+
+For Docker or Kubernetes deployments, prefer a private scrape path: bind the
+metrics listener to an internal interface and protect it with private
+networking, allowlists, NetworkPolicy, or proxy authentication. If you bind
+`METRICS_ADDR=0.0.0.0:9090` inside a container, only publish that port to a
+trusted network, for example a host-local mapping such as
+`127.0.0.1:9090:9090`.
+
+## Upgrading
+
+```bash
+docker compose -f docker-compose.selfhost.yml pull
+docker compose -f docker-compose.selfhost.yml up -d
+```
+
+Pin `MULTICA_IMAGE_TAG` in `.env` to an exact release like `v0.2.4` if you want to stay on a specific version. Migrations run automatically on backend startup. They are idempotent — running them multiple times has no effect.
+If the selected GHCR tag has not been published yet, fall back to `docker compose -f docker-compose.selfhost.yml -f docker-compose.selfhost.build.yml up -d --build`.

SELF_HOSTING_AI.md

@@ -0,0 +1,76 @@
+# Self-Hosting Setup (for AI Agents)
+
+This document is designed for AI agents to execute. Follow these steps exactly to deploy a local Multica instance and connect to it.
+
+## Prerequisites
+
+- Docker and Docker Compose installed
+- Homebrew installed (for CLI)
+- At least one AI agent CLI on PATH: `claude` or `codex`
+
+## Install
+
+```bash
+# Install CLI + provision self-host server
+curl -fsSL https://raw.githubusercontent.com/multica-ai/multica/main/scripts/install.sh | bash -s -- --with-server
+
+# Configure CLI for localhost, authenticate, and start daemon
+multica setup self-host
+```
+
+Wait for the server output `✓ Multica server is running and CLI is ready!` before running `multica setup self-host`.
+
+**Expected result:**
+- Frontend at http://localhost:3000
+- Backend at http://localhost:8080
+- `multica` CLI installed and configured for localhost
+
+## Alternative: Manual Setup
+
+```bash
+git clone https://github.com/multica-ai/multica.git
+cd multica
+make selfhost
+brew install multica-ai/tap/multica
+multica setup self-host
+```
+
+The `multica setup self-host` command will:
+1. Configure CLI to connect to localhost:8080 / localhost:3000
+2. Open a browser for login — use the emailed code, or the generated code printed in backend logs when Resend is unset
+3. Discover workspaces automatically
+4. Start the daemon in the background
+
+## Verification
+
+```bash
+multica daemon status
+```
+
+Should show `running` with detected agents.
+
+## Stopping
+
+```bash
+# Stop the daemon
+multica daemon stop
+
+# Stop all Docker services
+cd multica
+make selfhost-stop
+```
+
+## Custom Ports
+
+If the default ports (8080/3000) are in use:
+
+1. Edit `.env` and change `PORT` and `FRONTEND_PORT`
+2. Run `make selfhost`
+3. Run `multica setup self-host --port <PORT> --frontend-port <FRONTEND_PORT>`
+
+## Troubleshooting
+
+- **Backend not ready:** `docker compose -f docker-compose.selfhost.yml logs backend`
+- **Frontend not ready:** `docker compose -f docker-compose.selfhost.yml logs frontend`
+- **Daemon issues:** `multica daemon logs`
+- **Health checks:** `curl http://localhost:8080/health` for liveness, `curl http://localhost:8080/readyz` for dependency-aware readiness

apps/desktop/.gitignore

@@ -0,0 +1,8 @@
+node_modules
+dist
+out
+.DS_Store
+.eslintcache
+*.log*
+# CLI binary bundled at build time (from server/bin/)
+resources/bin/

apps/desktop/build/entitlements.mac.plist

@@ -0,0 +1,24 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
+<plist version="1.0">
+<dict>
+  <!-- Electron / V8 need JIT and unsigned executable memory under the
+       hardened runtime. -->
+  <key>com.apple.security.cs.allow-jit</key>
+  <true/>
+  <key>com.apple.security.cs.allow-unsigned-executable-memory</key>
+  <true/>
+  <!-- Required so the app can spawn the bundled `multica` Go binary and
+       any other child processes (e.g. agent CLIs) without Gatekeeper
+       blocking exec. -->
+  <key>com.apple.security.cs.disable-library-validation</key>
+  <true/>
+  <key>com.apple.security.cs.allow-dyld-environment-variables</key>
+  <true/>
+  <!-- Network client — the daemon talks to the backend + GitHub releases. -->
+  <key>com.apple.security.network.client</key>
+  <true/>
+  <key>com.apple.security.network.server</key>
+  <true/>
+</dict>
+</plist>

apps/desktop/electron-builder.yml

@@ -0,0 +1,62 @@
+appId: ai.multica.desktop
+productName: Multica
+directories:
+  buildResources: build
+files:
+  - "!**/.vscode/*"
+  - "!src/*"
+  - "!electron.vite.config.*"
+  - "!{.eslintignore,.eslintrc.cjs,.prettierignore,.prettierrc.yaml,dev-app-update.yml,CHANGELOG.md,README.md}"
+  - "!{tsconfig.json,tsconfig.node.json,tsconfig.web.json}"
+protocols:
+  - name: Multica
+    schemes:
+      - multica
+asarUnpack:
+  - resources/**
+mac:
+  entitlementsInherit: build/entitlements.mac.plist
+  target:
+    - dmg
+    - zip
+  # Hardcoded name avoids the `@multica/desktop-*` subdirectory that
+  # `${name}` produces for scoped package names.
+  # Naming scheme: multica-desktop-<version>-<platform>-<arch>.<ext>
+  # so the filename alone surfaces kind, version, platform, and CPU arch.
+  artifactName: multica-desktop-${version}-mac-${arch}.${ext}
+  # Notarize via notarytool. Requires APPLE_ID + APPLE_APP_SPECIFIC_PASSWORD
+  # + APPLE_TEAM_ID env vars at package time. Non-mac contributors are
+  # unaffected because `pnpm package` already requires the Developer ID
+  # signing cert — notarization is a strict superset.
+  notarize: true
+dmg:
+  artifactName: multica-desktop-${version}-mac-${arch}.${ext}
+linux:
+  target:
+    - AppImage
+    - deb
+    - rpm
+  artifactName: multica-desktop-${version}-linux-${arch}.${ext}
+rpm:
+  # Disable RPM build-id symlinks. Electron apps embed the upstream Electron
+  # binary, whose GNU build-id is identical across every app shipping the same
+  # Electron version (Slack, VS Code, Discord, ...). Without this, our RPM
+  # would own /usr/lib/.build-id/<hash> paths and collide with any other
+  # Electron RPM already installed, breaking `dnf install` on Fedora/RHEL.
+  fpm:
+    - "--rpm-rpmbuild-define=_build_id_links none"
+win:
+  target:
+    - nsis
+  artifactName: multica-desktop-${version}-windows-${arch}.${ext}
+publish:
+  provider: github
+  owner: multica-ai
+  repo: multica
+  # Align with our CLI release flow which pre-creates a *published* GitHub
+  # Release via `gh release create`. The electron-builder default of
+  # `releaseType: draft` conflicts with `existingType=release` and causes
+  # uploads of the DMG/ZIP/blockmaps/latest-mac.yml to be silently skipped,
+  # which breaks electron-updater auto-update on installed clients.
+  releaseType: release
+npmRebuild: false

apps/desktop/electron.vite.config.ts

@@ -0,0 +1,29 @@
+import { resolve } from "path";
+import { defineConfig, externalizeDepsPlugin } from "electron-vite";
+import react from "@vitejs/plugin-react";
+import tailwindcss from "@tailwindcss/vite";
+
+export default defineConfig({
+  main: {
+    plugins: [externalizeDepsPlugin()],
+  },
+  preload: {
+    plugins: [externalizeDepsPlugin()],
+  },
+  renderer: {
+    server: {
+      // Allow parallel worktrees to run `pnpm dev:desktop` side-by-side
+      // (e.g. Multica Canary alongside a primary checkout) by overriding
+      // the renderer port via env. Falls back to 5173 for the common case.
+      port: Number(process.env.DESKTOP_RENDERER_PORT) || 5173,
+      strictPort: true,
+    },
+    plugins: [react(), tailwindcss()],
+    resolve: {
+      alias: {
+        "@": resolve("src/renderer/src"),
+      },
+      dedupe: ["react", "react-dom"],
+    },
+  },
+});

apps/desktop/eslint.config.mjs

@@ -0,0 +1,37 @@
+import globals from "globals";
+import reactConfig from "@multica/eslint-config/react";
+
+export default [
+  ...reactConfig,
+  { ignores: ["out/", "dist/"] },
+  {
+    files: ["scripts/**/*.{mjs,js}"],
+    languageOptions: {
+      globals: { ...globals.node },
+    },
+  },
+  // Security: every renderer-controlled URL that reaches the OS shell must
+  // flow through openExternalSafely in src/main/external-url.ts (scheme
+  // allowlist). Enforce it statically so a direct shell.openExternal call
+  // cannot silently regress the protection.
+  {
+    files: ["src/main/**/*.ts"],
+    rules: {
+      "no-restricted-syntax": [
+        "error",
+        {
+          selector:
+            "CallExpression[callee.object.name='shell'][callee.property.name='openExternal']",
+          message:
+            "Do not call shell.openExternal directly. Use openExternalSafely from './external-url' so the http/https allowlist stays enforced.",
+        },
+      ],
+    },
+  },
+  {
+    files: ["src/main/external-url.ts"],
+    rules: {
+      "no-restricted-syntax": "off",
+    },
+  },
+];

apps/desktop/package.json

@@ -0,0 +1,74 @@
+{
+  "name": "@multica/desktop",
+  "version": "0.1.0",
+  "private": true,
+  "description": "Multica Desktop — native desktop client for the Multica platform.",
+  "homepage": "https://multica.ai",
+  "repository": {
+    "type": "git",
+    "url": "https://github.com/multica-ai/multica.git",
+    "directory": "apps/desktop"
+  },
+  "author": {
+    "name": "Multica",
+    "email": "support@multica.ai"
+  },
+  "license": "UNLICENSED",
+  "main": "./out/main/index.js",
+  "scripts": {
+    "bundle-cli": "node scripts/bundle-cli.mjs",
+    "brand-dev-electron": "node scripts/brand-dev-electron.mjs",
+    "dev": "pnpm run bundle-cli && pnpm run brand-dev-electron && electron-vite dev",
+    "dev:staging": "pnpm run bundle-cli && pnpm run brand-dev-electron && electron-vite dev --mode staging",
+    "build": "pnpm run bundle-cli && electron-vite build",
+    "typecheck:node": "tsc --noEmit -p tsconfig.node.json --composite false",
+    "typecheck:web": "tsc --noEmit -p tsconfig.web.json --composite false",
+    "typecheck": "pnpm run typecheck:node && pnpm run typecheck:web",
+    "preview": "electron-vite preview",
+    "package": "node scripts/package.mjs",
+    "package:all": "node scripts/package.mjs --all-platforms --publish never",
+    "lint": "eslint .",
+    "test": "vitest run",
+    "postinstall": "electron-builder install-app-deps"
+  },
+  "dependencies": {
+    "@dnd-kit/core": "^6.3.1",
+    "@dnd-kit/modifiers": "^9.0.0",
+    "@dnd-kit/sortable": "^10.0.0",
+    "@dnd-kit/utilities": "^3.2.2",
+    "@electron-toolkit/preload": "^3.0.2",
+    "@electron-toolkit/utils": "^4.0.0",
+    "@fontsource-variable/inter": "^5.2.5",
+    "@fontsource-variable/source-serif-4": "^5.2.9",
+    "@fontsource/geist-mono": "^5.2.7",
+    "@multica/core": "workspace:*",
+    "@multica/ui": "workspace:*",
+    "@multica/views": "workspace:*",
+    "electron-updater": "^6.8.3",
+    "fix-path": "^5.0.0",
+    "react-router-dom": "^7.6.0",
+    "shadcn": "^4.1.0",
+    "sonner": "^2.0.7",
+    "tw-animate-css": "^1.4.0"
+  },
+  "devDependencies": {
+    "@electron-toolkit/tsconfig": "^2.0.0",
+    "@multica/tsconfig": "workspace:*",
+    "@tailwindcss/vite": "^4",
+    "@testing-library/jest-dom": "catalog:",
+    "@testing-library/react": "catalog:",
+    "@types/node": "catalog:",
+    "@types/react": "catalog:",
+    "@types/react-dom": "catalog:",
+    "@vitejs/plugin-react": "^5.1.1",
+    "electron": "^39.2.6",
+    "electron-builder": "^26.0.12",
+    "electron-vite": "^5.0.0",
+    "jsdom": "catalog:",
+    "react": "catalog:",
+    "react-dom": "catalog:",
+    "tailwindcss": "^4",
+    "typescript": "catalog:",
+    "vitest": "catalog:"
+  }
+}

apps/desktop/scripts/brand-dev-electron.mjs

@@ -0,0 +1,73 @@
+#!/usr/bin/env node
+// Rebrand the bundled Electron.app's Info.plist so `pnpm dev:desktop`
+// shows "Multica Canary" in the menu bar, Cmd+Tab switcher, and
+// Activity Monitor. On macOS these titles come from CFBundleName at
+// launch time — `app.setName()` cannot override them at runtime, so
+// patching the plist in node_modules is the only working fix.
+//
+// Idempotent: runs on every dev launch and no-ops once the plist already
+// matches. The patch is isolated to this worktree's node_modules — we
+// unlink the file before rewriting so we never mutate a pnpm-store inode
+// shared with another project.
+
+import { createRequire } from "node:module";
+import { execFileSync } from "node:child_process";
+import { readFileSync, unlinkSync, writeFileSync } from "node:fs";
+import { resolve } from "node:path";
+
+if (process.platform !== "darwin") process.exit(0);
+
+const DESIRED_NAME = "Multica Canary";
+
+const require = createRequire(import.meta.url);
+// `require('electron')` returns the path to the executable
+// (.../Electron.app/Contents/MacOS/Electron). Walk up to Contents/Info.plist.
+const electronBin = require("electron");
+const plistPath = resolve(electronBin, "../../Info.plist");
+
+function plistGet(key) {
+  try {
+    return execFileSync(
+      "/usr/libexec/PlistBuddy",
+      ["-c", `Print :${key}`, plistPath],
+      { encoding: "utf8", stdio: ["ignore", "pipe", "ignore"] },
+    ).trim();
+  } catch {
+    return "";
+  }
+}
+
+function plistSet(key, value) {
+  try {
+    execFileSync("/usr/libexec/PlistBuddy", [
+      "-c",
+      `Set :${key} ${value}`,
+      plistPath,
+    ]);
+  } catch {
+    execFileSync("/usr/libexec/PlistBuddy", [
+      "-c",
+      `Add :${key} string ${value}`,
+      plistPath,
+    ]);
+  }
+}
+
+if (
+  plistGet("CFBundleName") === DESIRED_NAME &&
+  plistGet("CFBundleDisplayName") === DESIRED_NAME
+) {
+  process.exit(0);
+}
+
+// Break any pnpm hardlink to the global store: read, unlink, rewrite.
+// PlistBuddy would otherwise write through the hardlink and mutate the
+// shared store file (and every other project's Electron.app with it).
+const original = readFileSync(plistPath);
+unlinkSync(plistPath);
+writeFileSync(plistPath, original);
+
+plistSet("CFBundleName", DESIRED_NAME);
+plistSet("CFBundleDisplayName", DESIRED_NAME);
+
+console.log(`[brand-dev-electron] ${plistPath} → CFBundleName="${DESIRED_NAME}"`);

apps/desktop/scripts/bundle-cli.mjs

@@ -0,0 +1,168 @@
+#!/usr/bin/env node
+// Builds the `multica` CLI from server/cmd/multica and copies the binary
+// into apps/desktop/resources/bin/ so electron-vite (dev) and electron-
+// builder (prod) pick it up. Running this on every dev/build/package
+// invocation guarantees the bundled CLI always matches the current Go
+// source — no more stale binary surprises. Go's build cache makes the
+// no-op case (nothing changed) effectively free.
+//
+// ldflags mirror `make build` so `multica --version` reports a meaningful
+// version / commit / date.
+//
+// Graceful: if `go` is not installed (e.g. frontend-only contributor), we
+// skip the build and fall through to auto-install at runtime. A genuine
+// Go compile error is fatal — you want that to block dev, not hide.
+
+import { access, chmod, copyFile, mkdir, rm } from "node:fs/promises";
+import { constants } from "node:fs";
+import { execFileSync, execSync } from "node:child_process";
+import { dirname, join, resolve } from "node:path";
+import { fileURLToPath } from "node:url";
+
+const here = dirname(fileURLToPath(import.meta.url));
+const repoRoot = resolve(here, "..", "..", "..");
+const serverDir = join(repoRoot, "server");
+
+const PLATFORM_TO_GOOS = {
+  darwin: "darwin",
+  linux: "linux",
+  win32: "windows",
+};
+
+const SUPPORTED_ARCHS = new Set(["x64", "arm64"]);
+
+function runtimePlatformFromArgs(argv) {
+  const flagIndex = argv.indexOf("--target-platform");
+  if (flagIndex === -1) return process.platform;
+  return argv[flagIndex + 1] ?? "";
+}
+
+function runtimeArchFromArgs(argv) {
+  const flagIndex = argv.indexOf("--target-arch");
+  if (flagIndex === -1) return process.arch;
+  return argv[flagIndex + 1] ?? "";
+}
+
+function normalizeRuntimePlatform(platform) {
+  if (platform in PLATFORM_TO_GOOS) return platform;
+  throw new Error(
+    `[bundle-cli] unsupported target platform: ${platform}. ` +
+      "Use darwin, linux, or win32.",
+  );
+}
+
+function normalizeRuntimeArch(arch) {
+  if (SUPPORTED_ARCHS.has(arch)) return arch;
+  throw new Error(
+    `[bundle-cli] unsupported target architecture: ${arch}. ` +
+      "Use x64 or arm64.",
+  );
+}
+
+function binaryNameForPlatform(platform) {
+  return platform === "win32" ? "multica.exe" : "multica";
+}
+
+const targetPlatform = normalizeRuntimePlatform(
+  runtimePlatformFromArgs(process.argv.slice(2)),
+);
+const targetArch = normalizeRuntimeArch(runtimeArchFromArgs(process.argv.slice(2)));
+const goos = PLATFORM_TO_GOOS[targetPlatform];
+const goarch = targetArch === "x64" ? "amd64" : targetArch;
+const binName = binaryNameForPlatform(targetPlatform);
+const srcBinary = join(serverDir, "bin", `${goos}-${goarch}`, binName);
+const destDir = join(repoRoot, "apps", "desktop", "resources", "bin");
+const destBinary = join(destDir, binName);
+
+function sh(cmd) {
+  try {
+    return execSync(cmd, { encoding: "utf-8" }).trim();
+  } catch {
+    return "";
+  }
+}
+
+function hasGo() {
+  try {
+    execSync("go version", { stdio: "pipe" });
+    return true;
+  } catch {
+    return false;
+  }
+}
+
+async function exists(p) {
+  try {
+    await access(p, constants.F_OK);
+    return true;
+  } catch {
+    return false;
+  }
+}
+
+if (hasGo()) {
+  const version = sh("git describe --tags --always --dirty") || "dev";
+  const commit = sh("git rev-parse --short HEAD") || "unknown";
+  const date = new Date().toISOString().replace(/\.\d+Z$/, "Z");
+  const ldflags = `-X main.version=${version} -X main.commit=${commit} -X main.date=${date}`;
+
+  console.log(
+    `[bundle-cli] go build → ${srcBinary} (${goos}/${goarch}, version=${version} commit=${commit})`,
+  );
+  await mkdir(join(serverDir, "bin", `${goos}-${goarch}`), { recursive: true });
+  execFileSync(
+    "go",
+    [
+      "build",
+      "-ldflags",
+      ldflags,
+      "-o",
+      srcBinary,
+      "./cmd/multica",
+    ],
+    {
+      cwd: serverDir,
+      stdio: "inherit",
+      env: {
+        ...process.env,
+        CGO_ENABLED: "0",
+        GOOS: goos,
+        GOARCH: goarch,
+      },
+    },
+  );
+} else {
+  console.warn(
+    "[bundle-cli] `go` not found in PATH — skipping CLI build. " +
+      "Desktop will use whatever is already in resources/bin/, or fall back " +
+      "to auto-installing the latest release at runtime.",
+  );
+}
+
+if (!(await exists(srcBinary))) {
+  console.warn(
+    `[bundle-cli] ${srcBinary} not present — Desktop will fall back to ` +
+      `auto-installing the latest release at runtime.`,
+  );
+  await rm(destDir, { recursive: true, force: true });
+  process.exit(0);
+}
+
+await rm(destDir, { recursive: true, force: true });
+await mkdir(destDir, { recursive: true });
+await copyFile(srcBinary, destBinary);
+await chmod(destBinary, 0o755);
+
+// macOS: ad-hoc sign so Gatekeeper doesn't complain when the parent app
+// (which itself may be unsigned in dev) spawns the child.
+if (process.platform === "darwin") {
+  try {
+    execSync(`codesign -s - --force ${JSON.stringify(destBinary)}`, {
+      stdio: "pipe",
+    });
+  } catch {
+    // Non-fatal. Unsigned binaries still run when the parent app is trusted.
+  }
+}
+
+console.log(`[bundle-cli] bundled ${srcBinary} → ${destBinary}`);

apps/desktop/scripts/package.mjs

@@ -0,0 +1,430 @@
+#!/usr/bin/env node
+// Wrapper around `electron-builder` that keeps the Desktop version in
+// lockstep with the CLI. Both are derived from `git describe --tags
+// --always --dirty` — the same source GoReleaser reads for the CLI
+// binary via the `main.version` ldflag — so a single `vX.Y.Z` tag push
+// produces matching CLI and Desktop versions.
+//
+// Builds the Electron bundles once, then for each requested target
+// (platform + arch) compiles the matching Go CLI into resources/bin/ and
+// invokes electron-builder with `-c.extraMetadata.version=<derived>` so
+// the override applies at build time without mutating the tracked
+// package.json.
+//
+// The electron-vite step is important: electron-builder only packages
+// whatever is already in out/, so skipping it (or relying on stale
+// artifacts from a prior partial build) ships an app with missing
+// renderer code and white-screens on launch.
+//
+// Extra CLI args after `pnpm package --` are forwarded to electron-builder
+// unchanged (e.g. `--mac --arm64`). For an unsigned local smoke-test
+// build, set `CSC_IDENTITY_AUTO_DISCOVERY=false` so electron-builder falls
+// back to an ad-hoc signature instead of requiring a Developer ID cert.
+//
+// The `normalizeGitVersion` helper is exported so tests can cover the
+// version-derivation logic without shelling out.
+
+import { execFileSync, spawnSync, execSync } from "node:child_process";
+import { delimiter, dirname, resolve } from "node:path";
+import { fileURLToPath, pathToFileURL } from "node:url";
+
+const here = dirname(fileURLToPath(import.meta.url));
+const desktopRoot = resolve(here, "..");
+const bundleCliScript = resolve(here, "bundle-cli.mjs");
+
+const PLATFORM_CONFIG = {
+  mac: {
+    aliases: new Set(["--mac", "--macos", "-m"]),
+    builderFlag: "--mac",
+    runtimePlatform: "darwin",
+    label: "macOS",
+  },
+  win: {
+    aliases: new Set(["--win", "--windows", "-w"]),
+    builderFlag: "--win",
+    runtimePlatform: "win32",
+    label: "Windows",
+  },
+  linux: {
+    aliases: new Set(["--linux", "-l"]),
+    builderFlag: "--linux",
+    runtimePlatform: "linux",
+    label: "Linux",
+  },
+};
+
+const ARCH_FLAGS = new Map([
+  ["--x64", "x64"],
+  ["--arm64", "arm64"],
+  ["--ia32", "ia32"],
+  ["--armv7l", "armv7l"],
+  ["--universal", "universal"],
+]);
+
+const SUPPORTED_CLI_ARCHS = new Set(["x64", "arm64"]);
+const MAC_ALL_PLATFORM_TARGETS = [
+  { platform: "mac", arch: "arm64" },
+  { platform: "win", arch: "x64" },
+  { platform: "win", arch: "arm64" },
+  { platform: "linux", arch: "x64" },
+  { platform: "linux", arch: "arm64" },
+];
+
+function sh(cmd) {
+  try {
+    return execSync(cmd, { encoding: "utf-8" }).trim();
+  } catch {
+    return "";
+  }
+}
+
+/**
+ * Strip the leading `--` that npm/pnpm insert to separate their own
+ * flags from the ones meant for the underlying script.  Without this,
+ * `pnpm package -- --mac --arm64 --publish always` forwards the bare
+ * `--` into electron-builder's argv, which terminates option parsing
+ * and turns `--publish always` into ignored positional arguments.
+ */
+export function stripLeadingSeparator(argv) {
+  if (argv.length > 0 && argv[0] === "--") return argv.slice(1);
+  return argv;
+}
+
+/**
+ * Pure transformation from the `git describe --tags --always --dirty`
+ * output to the value we feed into electron-builder's extraMetadata.version.
+ *
+ *   - empty input              → null   (caller should fall back)
+ *   - "v0.1.36"                → "0.1.36"
+ *   - "v0.1.35-14-gf1415e96"   → "0.1.35-14-gf1415e96"  (semver prerelease)
+ *   - "v0.1.35-…-dirty"        → same, dirty suffix preserved
+ *   - "f1415e96" (no tag)      → "0.0.0-f1415e96"        (fallback)
+ *
+ * Leading `v` is stripped so the result is valid semver for package.json.
+ */
+export function normalizeGitVersion(raw) {
+  if (!raw) return null;
+  const stripped = raw.replace(/^v/, "");
+  if (!/^\d/.test(stripped)) {
+    // No reachable tag — `git describe` fell back to just the commit hash.
+    return `0.0.0-${stripped}`;
+  }
+  return stripped;
+}
+
+function deriveVersion() {
+  return normalizeGitVersion(sh("git describe --tags --always --dirty"));
+}
+
+function uniqueOrdered(values) {
+  return [...new Set(values)];
+}
+
+export function envWithLocalBins(env = process.env, root = desktopRoot) {
+  const pathKey =
+    Object.keys(env).find((key) => key.toUpperCase() === "PATH") ?? "PATH";
+  const existingPath = env[pathKey] ?? "";
+  const localBins = uniqueOrdered([
+    resolve(root, "node_modules", ".bin"),
+    resolve(root, "..", "..", "node_modules", ".bin"),
+  ]);
+  const mergedPath = uniqueOrdered([
+    ...localBins,
+    ...String(existingPath)
+      .split(delimiter)
+      .filter(Boolean),
+  ]).join(delimiter);
+  return { ...env, [pathKey]: mergedPath };
+}
+
+function hostPlatformKey(platform = process.platform) {
+  if (platform === "darwin") return "mac";
+  if (platform === "win32") return "win";
+  if (platform === "linux") return "linux";
+  throw new Error(`[package] unsupported host platform: ${platform}`);
+}
+
+function hostArchKey(arch = process.arch) {
+  if (SUPPORTED_CLI_ARCHS.has(arch)) return arch;
+  throw new Error(
+    `[package] unsupported host architecture for Desktop CLI bundling: ${arch}`,
+  );
+}
+
+function expandPlatformShorthand(token) {
+  if (!/^-[mwl]{2,}$/.test(token)) return null;
+  const expanded = [];
+  for (const char of token.slice(1)) {
+    if (char === "m") expanded.push("mac");
+    if (char === "w") expanded.push("win");
+    if (char === "l") expanded.push("linux");
+  }
+  return uniqueOrdered(expanded);
+}
+
+function platformKeyForToken(token) {
+  for (const [platform, config] of Object.entries(PLATFORM_CONFIG)) {
+    if (config.aliases.has(token)) return platform;
+  }
+  return null;
+}
+
+function platformTargetsTemplate() {
+  return { mac: [], win: [], linux: [] };
+}
+
+export function parsePackageArgs(argv) {
+  const sharedArgs = [];
+  const platformTargets = platformTargetsTemplate();
+  const requestedPlatforms = [];
+  const requestedArchs = [];
+  let allPlatforms = false;
+
+  for (let i = 0; i < argv.length; i += 1) {
+    const token = argv[i];
+    if (token === "--all-platforms") {
+      allPlatforms = true;
+      continue;
+    }
+
+    const expandedPlatforms = expandPlatformShorthand(token);
+    if (expandedPlatforms) {
+      requestedPlatforms.push(...expandedPlatforms);
+      continue;
+    }
+
+    const platform = platformKeyForToken(token);
+    if (platform) {
+      requestedPlatforms.push(platform);
+      while (i + 1 < argv.length && !argv[i + 1].startsWith("-")) {
+        platformTargets[platform].push(argv[i + 1]);
+        i += 1;
+      }
+      continue;
+    }
+
+    const arch = ARCH_FLAGS.get(token);
+    if (arch) {
+      requestedArchs.push(arch);
+      continue;
+    }
+
+    sharedArgs.push(token);
+  }
+
+  return {
+    allPlatforms,
+    sharedArgs,
+    platformTargets,
+    requestedPlatforms: uniqueOrdered(requestedPlatforms),
+    requestedArchs: uniqueOrdered(requestedArchs),
+  };
+}
+
+export function resolveBuildMatrix(parsed, platform = process.platform, arch = process.arch) {
+  if (parsed.allPlatforms) {
+    if (parsed.requestedPlatforms.length > 0 || parsed.requestedArchs.length > 0) {
+      throw new Error(
+        "[package] --all-platforms cannot be combined with explicit platform or arch flags",
+      );
+    }
+    if (platform !== "darwin") {
+      throw new Error(
+        `[package] --all-platforms is only supported on macOS hosts (current: ${platform})`,
+      );
+    }
+    return MAC_ALL_PLATFORM_TARGETS.map((target) => ({ ...target }));
+  }
+
+  const platforms =
+    parsed.requestedPlatforms.length > 0
+      ? parsed.requestedPlatforms
+      : [hostPlatformKey(platform)];
+  const archs =
+    parsed.requestedArchs.length > 0
+      ? parsed.requestedArchs
+      : [hostArchKey(arch)];
+
+  const unsupported = archs.filter((value) => !SUPPORTED_CLI_ARCHS.has(value));
+  if (unsupported.length > 0) {
+    throw new Error(
+      `[package] unsupported Desktop CLI architecture(s): ${unsupported.join(", ")}. ` +
+        "Use --x64 or --arm64.",
+    );
+  }
+
+  return platforms.flatMap((targetPlatform) =>
+    archs.map((targetArch) => ({
+      platform: targetPlatform,
+      arch: targetArch,
+    })),
+  );
+}
+
+function formatTarget(target) {
+  return `${PLATFORM_CONFIG[target.platform].label} ${target.arch}`;
+}
+
+export function builderArgsForTarget(
+  target,
+  parsed,
+  version,
+  {
+    disableMacNotarize = false,
+    hostPlatform = process.platform,
+    useScopedOutputDir = false,
+  } = {},
+) {
+  const builderArgs = [];
+  if (version) builderArgs.push(`-c.extraMetadata.version=${version}`);
+  if (disableMacNotarize) builderArgs.push("-c.mac.notarize=false");
+  builderArgs.push(PLATFORM_CONFIG[target.platform].builderFlag);
+  const requestedTargets = parsed.platformTargets[target.platform];
+  if (
+    target.platform === "linux" &&
+    hostPlatform !== "linux" &&
+    requestedTargets.length === 0
+  ) {
+    // electron-builder only guarantees AppImage/Snap when cross-building
+    // Linux from macOS/Windows. Keep `package:all` portable by defaulting
+    // to AppImage unless the caller explicitly requests Linux targets.
+    builderArgs.push("AppImage");
+  } else {
+    builderArgs.push(...requestedTargets);
+  }
+  builderArgs.push(`--${target.arch}`);
+  builderArgs.push(...parsed.sharedArgs);
+  if (useScopedOutputDir) {
+    builderArgs.push(
+      `-c.directories.output=dist/${target.platform}-${target.arch}`,
+    );
+  }
+  // electron-builder's update metadata file is `latest.yml` for Windows
+  // regardless of arch (only Linux gets an arch suffix automatically — see
+  // app-builder-lib's getArchPrefixForUpdateFile). Without an explicit
+  // channel override, building Windows x64 and arm64 in two invocations
+  // makes both publish `latest.yml` to the same GitHub Release, so the
+  // second upload overwrites the first and one of the two architectures
+  // ends up with no auto-update metadata. Route Windows arm64 to its own
+  // channel so x64 keeps `latest.yml` and arm64 ships `latest-arm64.yml`;
+  // the renderer-side updater pins the matching channel per arch.
+  if (target.platform === "win" && target.arch === "arm64") {
+    builderArgs.push("-c.publish.channel=latest-arm64");
+  }
+  return builderArgs;
+}
+
+function main() {
+  const passthrough = stripLeadingSeparator(process.argv.slice(2));
+  const parsed = parsePackageArgs(passthrough);
+  const buildMatrix = resolveBuildMatrix(parsed);
+  console.log(
+    `[package] build matrix → ${buildMatrix.map(formatTarget).join(", ")}`,
+  );
+
+  // Step 1: build the Electron main/preload/renderer bundles. Without
+  // this step electron-builder silently packages whatever is already in
+  // out/, which on a fresh checkout (or after a partial build) ships an
+  // app that white-screens because the renderer bundle is missing.
+  //
+  // CI invokes this script via `node scripts/package.mjs`, so we cannot
+  // rely on pnpm/npm to inject package-local binaries into PATH.
+  //
+  // `shell: true` is required on Windows: `node_modules/.bin/electron-vite`
+  // ships as a `.cmd` shim there, and Node's `spawnSync` does not honour
+  // PATHEXT when spawning a bare command without a shell — it would fail
+  // with `ENOENT`. On POSIX hosts the shim is a real executable so going
+  // through the shell is harmless. See
+  // https://nodejs.org/api/child_process.html#spawning-bat-and-cmd-files-on-windows
+  const viteResult = spawnSync("electron-vite", ["build"], {
+    stdio: "inherit",
+    cwd: desktopRoot,
+    env: envWithLocalBins(),
+    shell: true,
+  });
+  if (viteResult.error) {
+    console.error(
+      "[package] failed to spawn electron-vite:",
+      viteResult.error.message,
+    );
+    process.exit(1);
+  }
+  if (viteResult.status !== 0) {
+    process.exit(viteResult.status ?? 1);
+  }
+
+  // Step 2: derive the version that should be written into the app.
+  const version = deriveVersion();
+  if (version) {
+    console.log(`[package] Desktop version → ${version} (from git describe)`);
+  } else {
+    console.warn(
+      "[package] could not derive version from git; falling back to package.json",
+    );
+  }
+
+  const disableMacNotarize = !process.env.APPLE_TEAM_ID;
+  if (disableMacNotarize) {
+    console.warn(
+      "[package] APPLE_TEAM_ID not set — skipping notarization (local dev build). " +
+        "Set APPLE_ID + APPLE_APP_SPECIFIC_PASSWORD + APPLE_TEAM_ID for a release build.",
+    );
+  }
+
+  const useScopedOutputDir = buildMatrix.length > 1;
+
+  // Step 3: for each requested target, build the matching CLI into
+  // resources/bin/ and package that target in isolation.
+  for (const target of buildMatrix) {
+    console.log(`[package] bundling CLI → ${formatTarget(target)}`);
+    execFileSync(
+      "node",
+      [
+        bundleCliScript,
+        "--target-platform",
+        PLATFORM_CONFIG[target.platform].runtimePlatform,
+        "--target-arch",
+        target.arch,
+      ],
+      {
+        stdio: "inherit",
+        cwd: desktopRoot,
+      },
+    );
+
+    const builderArgs = builderArgsForTarget(target, parsed, version, {
+      disableMacNotarize,
+      hostPlatform: process.platform,
+      useScopedOutputDir,
+    });
+
+    // Step 4: invoke electron-builder for the current target only.
+    // `shell: true` for the same Windows `.cmd` shim reason as the
+    // electron-vite invocation above.
+    const result = spawnSync("electron-builder", builderArgs, {
+      stdio: "inherit",
+      cwd: desktopRoot,
+      env: envWithLocalBins(),
+      shell: true,
+    });
+
+    if (result.error) {
+      console.error(
+        "[package] failed to spawn electron-builder:",
+        result.error.message,
+      );
+      process.exit(1);
+    }
+    if (result.status !== 0) {
+      process.exit(result.status ?? 1);
+    }
+  }
+}
+
+// Only run when invoked as a CLI, not when imported by a test file.
+if (
+  process.argv[1] &&
+  import.meta.url === pathToFileURL(process.argv[1]).href
+) {
+  main();
+}

apps/desktop/scripts/package.test.mjs

@@ -0,0 +1,273 @@
+import { delimiter, resolve } from "node:path";
+import { describe, it, expect } from "vitest";
+import {
+  builderArgsForTarget,
+  envWithLocalBins,
+  normalizeGitVersion,
+  parsePackageArgs,
+  resolveBuildMatrix,
+  stripLeadingSeparator,
+} from "./package.mjs";
+
+describe("normalizeGitVersion", () => {
+  it("returns null for empty / nullish input", () => {
+    expect(normalizeGitVersion("")).toBe(null);
+    expect(normalizeGitVersion(null)).toBe(null);
+    expect(normalizeGitVersion(undefined)).toBe(null);
+  });
+
+  it("strips the leading v on a clean tag", () => {
+    expect(normalizeGitVersion("v0.1.36")).toBe("0.1.36");
+    expect(normalizeGitVersion("v1.0.0")).toBe("1.0.0");
+  });
+
+  it("preserves the prerelease suffix between tags", () => {
+    expect(normalizeGitVersion("v0.1.35-14-gf1415e96")).toBe(
+      "0.1.35-14-gf1415e96",
+    );
+  });
+
+  it("preserves the dirty suffix on a modified worktree", () => {
+    expect(normalizeGitVersion("v0.1.35-14-gf1415e96-dirty")).toBe(
+      "0.1.35-14-gf1415e96-dirty",
+    );
+  });
+
+  it("handles v-prefixed prerelease tags", () => {
+    expect(normalizeGitVersion("v1.0.0-alpha")).toBe("1.0.0-alpha");
+    expect(normalizeGitVersion("v1.0.0-rc.2")).toBe("1.0.0-rc.2");
+  });
+
+  it("falls back to 0.0.0-<hash> when no tags are reachable", () => {
+    // `git describe --tags --always` returns just the short commit hash
+    // when there are no tags in the history at all.
+    expect(normalizeGitVersion("f1415e96")).toBe("0.0.0-f1415e96");
+    expect(normalizeGitVersion("abc1234")).toBe("0.0.0-abc1234");
+  });
+});
+
+describe("stripLeadingSeparator", () => {
+  it("removes the leading -- inserted by npm/pnpm", () => {
+    expect(stripLeadingSeparator(["--", "--mac", "--arm64", "--publish", "always"])).toEqual([
+      "--mac", "--arm64", "--publish", "always",
+    ]);
+  });
+
+  it("leaves args untouched when there is no leading --", () => {
+    expect(stripLeadingSeparator(["--mac", "--arm64"])).toEqual(["--mac", "--arm64"]);
+  });
+
+  it("does not strip a -- that appears mid-argv", () => {
+    expect(stripLeadingSeparator(["--mac", "--", "--arm64"])).toEqual([
+      "--mac", "--", "--arm64",
+    ]);
+  });
+
+  it("handles an empty array", () => {
+    expect(stripLeadingSeparator([])).toEqual([]);
+  });
+});
+
+describe("parsePackageArgs", () => {
+  it("collects per-platform targets and shared args", () => {
+    expect(
+      parsePackageArgs([
+        "--win", "nsis",
+        "--mac", "dmg", "zip",
+        "--arm64",
+        "--publish", "never",
+      ]),
+    ).toEqual({
+      allPlatforms: false,
+      sharedArgs: ["--publish", "never"],
+      platformTargets: {
+        mac: ["dmg", "zip"],
+        win: ["nsis"],
+        linux: [],
+      },
+      requestedPlatforms: ["win", "mac"],
+      requestedArchs: ["arm64"],
+    });
+  });
+
+  it("expands combined short flags", () => {
+    expect(parsePackageArgs(["-mw", "--x64"]).requestedPlatforms).toEqual([
+      "mac",
+      "win",
+    ]);
+  });
+
+  it("tracks the all-platforms shortcut", () => {
+    expect(parsePackageArgs(["--all-platforms", "--publish", "never"]).allPlatforms).toBe(true);
+  });
+});
+
+describe("resolveBuildMatrix", () => {
+  it("defaults to the current host platform and arch", () => {
+    expect(
+      resolveBuildMatrix(
+        {
+          allPlatforms: false,
+          sharedArgs: [],
+          platformTargets: { mac: [], win: [], linux: [] },
+          requestedPlatforms: [],
+          requestedArchs: [],
+        },
+        "darwin",
+        "arm64",
+      ),
+    ).toEqual([{ platform: "mac", arch: "arm64" }]);
+  });
+
+  it("expands all-platforms on macOS", () => {
+    expect(
+      resolveBuildMatrix(
+        {
+          allPlatforms: true,
+          sharedArgs: [],
+          platformTargets: { mac: [], win: [], linux: [] },
+          requestedPlatforms: [],
+          requestedArchs: [],
+        },
+        "darwin",
+        "arm64",
+      ),
+    ).toEqual([
+      { platform: "mac", arch: "arm64" },
+      { platform: "win", arch: "x64" },
+      { platform: "win", arch: "arm64" },
+      { platform: "linux", arch: "x64" },
+      { platform: "linux", arch: "arm64" },
+    ]);
+  });
+
+  it("rejects unsupported architectures", () => {
+    expect(() =>
+      resolveBuildMatrix(
+        {
+          allPlatforms: false,
+          sharedArgs: [],
+          platformTargets: { mac: [], win: [], linux: [] },
+          requestedPlatforms: ["win"],
+          requestedArchs: ["universal"],
+        },
+        "darwin",
+        "arm64",
+      ),
+    ).toThrow(/unsupported Desktop CLI architecture/);
+  });
+});
+
+describe("builderArgsForTarget", () => {
+  it("adds scoped output directories for multi-target builds", () => {
+    expect(
+      builderArgsForTarget(
+        { platform: "win", arch: "arm64" },
+        {
+          allPlatforms: false,
+          sharedArgs: ["--publish", "never"],
+          platformTargets: { mac: [], win: ["nsis"], linux: [] },
+          requestedPlatforms: ["win"],
+          requestedArchs: ["arm64"],
+        },
+        "1.2.3",
+        {
+          disableMacNotarize: true,
+          hostPlatform: "darwin",
+          useScopedOutputDir: true,
+        },
+      ),
+    ).toEqual([
+      "-c.extraMetadata.version=1.2.3",
+      "-c.mac.notarize=false",
+      "--win",
+      "nsis",
+      "--arm64",
+      "--publish",
+      "never",
+      "-c.directories.output=dist/win-arm64",
+      "-c.publish.channel=latest-arm64",
+    ]);
+  });
+
+  it("does not override the publish channel for Windows x64 (default latest.yml)", () => {
+    expect(
+      builderArgsForTarget(
+        { platform: "win", arch: "x64" },
+        {
+          allPlatforms: false,
+          sharedArgs: ["--publish", "always"],
+          platformTargets: { mac: [], win: ["nsis"], linux: [] },
+          requestedPlatforms: ["win"],
+          requestedArchs: ["x64"],
+        },
+        "1.2.3",
+        { hostPlatform: "win32", useScopedOutputDir: true },
+      ),
+    ).toEqual([
+      "-c.extraMetadata.version=1.2.3",
+      "--win",
+      "nsis",
+      "--x64",
+      "--publish",
+      "always",
+      "-c.directories.output=dist/win-x64",
+    ]);
+  });
+
+  it("defaults linux cross-builds to AppImage on non-Linux hosts", () => {
+    expect(
+      builderArgsForTarget(
+        { platform: "linux", arch: "x64" },
+        {
+          allPlatforms: false,
+          sharedArgs: ["--publish", "never"],
+          platformTargets: { mac: [], win: [], linux: [] },
+          requestedPlatforms: ["linux"],
+          requestedArchs: ["x64"],
+        },
+        "1.2.3",
+        { hostPlatform: "darwin" },
+      ),
+    ).toEqual([
+      "-c.extraMetadata.version=1.2.3",
+      "--linux",
+      "AppImage",
+      "--x64",
+      "--publish",
+      "never",
+    ]);
+  });
+});
+
+describe("envWithLocalBins", () => {
+  it("prepends desktop-local binary directories to PATH", () => {
+    const desktopRoot = "/repo/apps/desktop";
+    const result = envWithLocalBins(
+      { PATH: ["/usr/local/bin", "/usr/bin"].join(delimiter) },
+      desktopRoot,
+    );
+    expect(result.PATH.split(delimiter)).toEqual([
+      resolve(desktopRoot, "node_modules", ".bin"),
+      resolve(desktopRoot, "..", "..", "node_modules", ".bin"),
+      "/usr/local/bin",
+      "/usr/bin",
+    ]);
+  });
+
+  it("preserves an existing Path key and avoids duplicate entries", () => {
+    const desktopRoot = "/repo/apps/desktop";
+    const desktopBin = resolve(desktopRoot, "node_modules", ".bin");
+    const workspaceBin = resolve(desktopRoot, "..", "..", "node_modules", ".bin");
+    const result = envWithLocalBins(
+      { Path: [desktopBin, "runner-bin", workspaceBin].join(delimiter) },
+      desktopRoot,
+    );
+    expect(result).not.toHaveProperty("PATH");
+    expect(result.Path.split(delimiter)).toEqual([
+      desktopBin,
+      workspaceBin,
+      "runner-bin",
+    ]);
+  });
+});

apps/desktop/src/main/app-version.ts

@@ -0,0 +1,33 @@
+import { app } from "electron";
+import { execSync } from "node:child_process";
+
+/**
+ * Resolve the running app version. In packaged builds this is the value
+ * `electron-builder` baked into package.json via `extraMetadata.version`
+ * (driven by `git describe` — see `apps/desktop/scripts/package.mjs`), so
+ * `app.getVersion()` matches the GitHub Release tag exactly.
+ *
+ * In dev (`pnpm dev:desktop`) `app.getVersion()` only sees the static
+ * `apps/desktop/package.json` value, which is "0.1.0" and never bumped —
+ * the Settings → Updates panel and any other UI surfacing the version
+ * would mislead developers into thinking they're running ancient builds.
+ * Fall back to `git describe --tags --always --dirty` (same source the
+ * packager uses) so dev shows e.g. `0.2.19-14-gabcdef-dirty`. If git is
+ * unavailable for whatever reason, we just return the package.json value.
+ */
+export function getAppVersion(): string {
+  if (app.isPackaged) {
+    return app.getVersion();
+  }
+  try {
+    const raw = execSync("git describe --tags --always --dirty", {
+      cwd: app.getAppPath(),
+      encoding: "utf-8",
+      stdio: ["ignore", "pipe", "ignore"],
+    }).trim();
+    if (!raw) return app.getVersion();
+    return raw.replace(/^v/, "");
+  } catch {
+    return app.getVersion();
+  }
+}

apps/desktop/src/main/cli-bootstrap.ts

@@ -0,0 +1,157 @@
+import { app } from "electron";
+import { execFile } from "child_process";
+import { createHash } from "crypto";
+import { createReadStream, createWriteStream, existsSync } from "fs";
+import { chmod, mkdir, rename, rm } from "fs/promises";
+import { join, dirname } from "path";
+import { pipeline } from "stream/promises";
+import { tmpdir } from "os";
+import { Readable } from "stream";
+
+import { selectPlatformReleaseAssetName } from "./cli-release-asset";
+
+// Desktop prefers the bundled `multica` CLI shipped inside the app for
+// same-repo builds, but it can also repair or bootstrap a managed copy in
+// userData on first launch when the bundled binary is missing or unusable.
+
+const GITHUB_LATEST_BASE =
+  "https://github.com/multica-ai/multica/releases/latest/download";
+
+function binaryName(): string {
+  return process.platform === "win32" ? "multica.exe" : "multica";
+}
+
+export function managedCliPath(): string {
+  return join(app.getPath("userData"), "bin", binaryName());
+}
+
+function run(cmd: string, args: string[], cwd?: string): Promise<void> {
+  return new Promise((resolve, reject) => {
+    execFile(cmd, args, { cwd }, (err) => (err ? reject(err) : resolve()));
+  });
+}
+
+async function downloadToFile(url: string, dest: string): Promise<void> {
+  const res = await fetch(url, { redirect: "follow" });
+  if (!res.ok || !res.body) {
+    throw new Error(`download failed: ${res.status} ${res.statusText}`);
+  }
+  await mkdir(dirname(dest), { recursive: true });
+  // Node's fetch returns a web ReadableStream; adapt to a Node stream for pipeline.
+  const nodeStream = Readable.fromWeb(res.body as Parameters<typeof Readable.fromWeb>[0]);
+  await pipeline(nodeStream, createWriteStream(dest));
+}
+
+// Fetch goreleaser's published checksums.txt and parse it into a
+// filename → sha256 lookup. Format is `<hex>  <filename>` per line.
+async function fetchChecksums(): Promise<Map<string, string>> {
+  const url = `${GITHUB_LATEST_BASE}/checksums.txt`;
+  const res = await fetch(url, { redirect: "follow" });
+  if (!res.ok) {
+    throw new Error(
+      `checksums.txt fetch failed: ${res.status} ${res.statusText}`,
+    );
+  }
+  const text = await res.text();
+  const map = new Map<string, string>();
+  for (const rawLine of text.split("\n")) {
+    const line = rawLine.trim();
+    if (!line) continue;
+    const match = line.match(/^([a-f0-9]{64})\s+\*?(\S+)$/i);
+    if (match) map.set(match[2], match[1].toLowerCase());
+  }
+  return map;
+}
+
+async function sha256OfFile(path: string): Promise<string> {
+  const hash = createHash("sha256");
+  await pipeline(createReadStream(path), hash);
+  return hash.digest("hex");
+}
+
+async function verifyChecksum(
+  archivePath: string,
+  assetName: string,
+  expected: string,
+): Promise<void> {
+  const actual = await sha256OfFile(archivePath);
+  if (actual.toLowerCase() !== expected) {
+    throw new Error(
+      `checksum mismatch for ${assetName}: expected ${expected}, got ${actual}`,
+    );
+  }
+}
+
+async function extractArchive(archive: string, dest: string): Promise<void> {
+  await mkdir(dest, { recursive: true });
+  // Modern OSes all ship a `tar` that auto-detects tar.gz and zip:
+  // - macOS/Linux: GNU tar or bsdtar
+  // - Windows 10+: bsdtar is bundled as `tar.exe` since build 17063
+  await run("tar", ["-xf", archive, "-C", dest]);
+}
+
+async function installFresh(): Promise<string> {
+  const target = managedCliPath();
+  const checksums = await fetchChecksums();
+  const assetName = selectPlatformReleaseAssetName(checksums.keys());
+  const expectedChecksum = checksums.get(assetName);
+  if (!expectedChecksum) {
+    throw new Error(
+      `no checksum for ${assetName} in checksums.txt — refusing to install unverified binary`,
+    );
+  }
+  const url = `${GITHUB_LATEST_BASE}/${assetName}`;
+
+  const workDir = join(tmpdir(), `multica-cli-${Date.now()}`);
+  await mkdir(workDir, { recursive: true });
+
+  try {
+    const archivePath = join(workDir, assetName);
+    console.log(`[cli-bootstrap] downloading ${url}`);
+    await downloadToFile(url, archivePath);
+
+    console.log(`[cli-bootstrap] verifying ${assetName} against checksums.txt`);
+    await verifyChecksum(archivePath, assetName, expectedChecksum);
+
+    console.log(`[cli-bootstrap] extracting ${assetName}`);
+    await extractArchive(archivePath, workDir);
+
+    const extractedBin = join(workDir, binaryName());
+    if (!existsSync(extractedBin)) {
+      throw new Error(
+        `archive ${assetName} did not contain ${binaryName()} at its root`,
+      );
+    }
+
+    await mkdir(dirname(target), { recursive: true });
+    await rm(target, { force: true }).catch(() => {});
+    await rename(extractedBin, target);
+    await chmod(target, 0o755);
+
+    // macOS: ad-hoc sign so spawning the child never hits a gatekeeper quirk.
+    // Non-fatal: unsigned binaries still execute when the parent app is trusted.
+    if (process.platform === "darwin") {
+      await run("codesign", ["-s", "-", "--force", target]).catch((err) => {
+        console.warn("[cli-bootstrap] ad-hoc codesign failed:", err);
+      });
+    }
+
+    console.log(`[cli-bootstrap] installed CLI at ${target}`);
+    return target;
+  } finally {
+    await rm(workDir, { recursive: true, force: true }).catch(() => {});
+  }
+}
+
+/**
+ * Returns the path to a usable `multica` binary. If one is already present at
+ * the managed userData location, returns it immediately. Otherwise downloads
+ * the latest release asset for the current platform and installs it.
+ */
+export async function ensureManagedCli(
+  options: { forceInstall?: boolean } = {},
+): Promise<string> {
+  const target = managedCliPath();
+  if (existsSync(target) && !options.forceInstall) return target;
+  return installFresh();
+}

apps/desktop/src/main/cli-release-asset.test.ts

@@ -0,0 +1,59 @@
+import { describe, expect, it } from "vitest";
+
+import { selectPlatformReleaseAssetName } from "./cli-release-asset";
+
+describe("selectPlatformReleaseAssetName", () => {
+  it("prefers the versioned archive name when both exist", () => {
+    const assetNames = [
+      "checksums.txt",
+      "multica_darwin_amd64.tar.gz",
+      "multica-cli-1.2.3-darwin-amd64.tar.gz",
+    ];
+
+    expect(selectPlatformReleaseAssetName(assetNames, "darwin", "x64")).toBe(
+      "multica-cli-1.2.3-darwin-amd64.tar.gz",
+    );
+  });
+
+  it("falls back to the legacy archive name when only legacy is present", () => {
+    const assetNames = ["checksums.txt", "multica_darwin_amd64.tar.gz"];
+
+    expect(selectPlatformReleaseAssetName(assetNames, "darwin", "x64")).toBe(
+      "multica_darwin_amd64.tar.gz",
+    );
+  });
+
+  it("matches the renamed darwin archive from release assets", () => {
+    const assetNames = [
+      "checksums.txt",
+      "multica-cli-1.2.3-darwin-amd64.tar.gz",
+      "multica-cli-1.2.3-darwin-arm64.tar.gz",
+      "multica-cli-1.2.3-linux-amd64.tar.gz",
+    ];
+
+    expect(selectPlatformReleaseAssetName(assetNames, "darwin", "x64")).toBe(
+      "multica-cli-1.2.3-darwin-amd64.tar.gz",
+    );
+  });
+
+  it("matches the renamed windows zip archive", () => {
+    const assetNames = [
+      "multica-cli-1.2.3-windows-amd64.zip",
+      "multica-cli-1.2.3-linux-amd64.tar.gz",
+    ];
+
+    expect(selectPlatformReleaseAssetName(assetNames, "win32", "x64")).toBe(
+      "multica-cli-1.2.3-windows-amd64.zip",
+    );
+  });
+
+  it("fails when the current platform asset is missing", () => {
+    expect(() =>
+      selectPlatformReleaseAssetName(
+        ["multica-cli-1.2.3-linux-amd64.tar.gz", "multica_linux_amd64.tar.gz"],
+        "darwin",
+        "arm64",
+      ),
+    ).toThrow(/no release asset found/);
+  });
+});

apps/desktop/src/main/cli-release-asset.ts

@@ -0,0 +1,62 @@
+const RELEASE_ARCHIVE_PREFIX = "multica-cli-";
+
+function platformArchiveDescriptor(
+  platform: NodeJS.Platform = process.platform,
+  arch: string = process.arch,
+): { os: string; arch: string; ext: string } {
+  const osMap: Record<string, string> = {
+    darwin: "darwin",
+    linux: "linux",
+    win32: "windows",
+  };
+  const archMap: Record<string, string> = {
+    x64: "amd64",
+    arm64: "arm64",
+  };
+  const os = osMap[platform];
+  const mappedArch = archMap[arch];
+  if (!os || !mappedArch) {
+    throw new Error(
+      `unsupported platform for CLI auto-install: ${platform}/${arch}`,
+    );
+  }
+  const ext = platform === "win32" ? "zip" : "tar.gz";
+  return { os, arch: mappedArch, ext };
+}
+
+export function selectPlatformReleaseAssetName(
+  assetNames: Iterable<string>,
+  platform: NodeJS.Platform = process.platform,
+  arch: string = process.arch,
+): string {
+  const { os, arch: mappedArch, ext } = platformArchiveDescriptor(
+    platform,
+    arch,
+  );
+  const names = [...assetNames];
+
+  // Prefer the versioned `multica-cli-<v>-<os>-<arch>.<ext>` name; fall
+  // back to the legacy `multica_<os>_<arch>.<ext>` so older releases that
+  // only ship the legacy archive keep working.
+  const suffix = `-${os}-${mappedArch}.${ext}`;
+  const matches = names.filter(
+    (name) =>
+      name.startsWith(RELEASE_ARCHIVE_PREFIX) && name.endsWith(suffix),
+  );
+
+  if (matches.length === 1) {
+    return matches[0];
+  }
+  if (matches.length > 1) {
+    throw new Error(
+      `multiple release assets matched current platform ${suffix}: ${matches.join(", ")}`,
+    );
+  }
+
+  const legacyName = `multica_${os}_${mappedArch}.${ext}`;
+  if (names.includes(legacyName)) {
+    return legacyName;
+  }
+
+  throw new Error(`no release asset found for current platform: ${suffix}`);
+}

apps/desktop/src/main/context-menu.ts

@@ -0,0 +1,33 @@
+import { BrowserWindow, Menu, MenuItem, type WebContents } from "electron";
+
+// Electron ships with no default right-click menu, so a user selecting text
+// in the renderer has no way to copy it. Mirror Chrome's minimal clipboard
+// menu using `roles`, which keeps i18n + accelerator handling native.
+export function installContextMenu(webContents: WebContents): void {
+  webContents.on("context-menu", (_event, params) => {
+    const { editFlags, selectionText, isEditable } = params;
+    const hasSelection = selectionText.trim().length > 0;
+
+    const menu = new Menu();
+
+    if (isEditable && editFlags.canCut) {
+      menu.append(new MenuItem({ role: "cut" }));
+    }
+    if (hasSelection && editFlags.canCopy) {
+      menu.append(new MenuItem({ role: "copy" }));
+    }
+    if (isEditable && editFlags.canPaste) {
+      menu.append(new MenuItem({ role: "paste" }));
+    }
+    if (isEditable && editFlags.canSelectAll) {
+      if (menu.items.length > 0) {
+        menu.append(new MenuItem({ type: "separator" }));
+      }
+      menu.append(new MenuItem({ role: "selectAll" }));
+    }
+
+    if (menu.items.length === 0) return;
+    const window = BrowserWindow.fromWebContents(webContents) ?? undefined;
+    menu.popup({ window });
+  });
+}

apps/desktop/src/main/daemon-manager.ts

@@ -0,0 +1,956 @@
+import { app, ipcMain, BrowserWindow, shell } from "electron";
+import { execFile } from "child_process";
+import {
+  readFile,
+  writeFile,
+  mkdir,
+  rm,
+  open,
+  stat,
+} from "fs/promises";
+import {
+  existsSync,
+  watchFile,
+  unwatchFile,
+  type StatsListener,
+} from "fs";
+import { join } from "path";
+import { homedir } from "os";
+import type { DaemonStatus, DaemonPrefs } from "../shared/daemon-types";
+import { ensureManagedCli, managedCliPath } from "./cli-bootstrap";
+import { decideVersionAction } from "./version-decision";
+
+const DEFAULT_HEALTH_PORT = 19514;
+const POLL_INTERVAL_MS = 5_000;
+const PREFS_PATH = join(homedir(), ".multica", "desktop_prefs.json");
+const LOG_TAIL_RETRY_MS = 2_000;
+const LOG_TAIL_MAX_RETRIES = 5;
+
+const DEFAULT_PREFS: DaemonPrefs = { autoStart: true, autoStop: false };
+
+interface ActiveProfile {
+  name: string; // "" = default profile
+  port: number;
+}
+
+let statusPollTimer: ReturnType<typeof setInterval> | null = null;
+let logTailWatcher: { path: string; listener: StatsListener } | null = null;
+let currentState: DaemonStatus["state"] = "installing_cli";
+let getMainWindow: () => BrowserWindow | null = () => null;
+let operationInProgress = false;
+let cachedCliBinary: string | null | undefined = undefined;
+let cliResolvePromise: Promise<string | null> | null = null;
+let cachedCliBinaryVersion: string | null | undefined = undefined;
+// Set when a CLI version mismatch was detected but the running daemon is
+// busy executing tasks. The poll loop retries the check on each tick and
+// fires the restart once active_task_count drops to 0.
+let pendingVersionRestart = false;
+let targetApiBaseUrl: string | null = null;
+let activeProfile: ActiveProfile | null = null;
+
+// Serialize all writes to any profile config file. Multiple paths
+// (syncToken, resolveActiveProfile, clearToken, watch/unwatch handlers)
+// may try to write concurrently; chaining them avoids interleaved writes
+// corrupting the JSON.
+let configWriteChain: Promise<void> = Promise.resolve();
+
+// Keep the Go impl in sync: server/cmd/multica/cmd_daemon.go healthPortForProfile.
+function healthPortForProfile(profile: string): number {
+  if (!profile) return DEFAULT_HEALTH_PORT;
+  let sum = 0;
+  for (const b of Buffer.from(profile, "utf-8")) sum += b;
+  return DEFAULT_HEALTH_PORT + 1 + (sum % 1000);
+}
+
+function profileDir(profile: string): string {
+  return profile
+    ? join(homedir(), ".multica", "profiles", profile)
+    : join(homedir(), ".multica");
+}
+
+function profileConfigPath(profile: string): string {
+  return join(profileDir(profile), "config.json");
+}
+
+function profileLogPath(profile: string): string {
+  return join(profileDir(profile), "daemon.log");
+}
+
+// Sidecar file that records which Multica user the cached PAT in config.json
+// was minted for. The Go CLI/daemon never read or write this file, so it
+// survives Go-side config rewrites. Used to detect user switches and mint a
+// fresh PAT instead of reusing a token that belongs to a previous user.
+function profileUserIdPath(profile: string): string {
+  return join(profileDir(profile), ".desktop-user-id");
+}
+
+async function readProfileUserId(profile: string): Promise<string | null> {
+  try {
+    const raw = await readFile(profileUserIdPath(profile), "utf-8");
+    const trimmed = raw.trim();
+    return trimmed || null;
+  } catch {
+    return null;
+  }
+}
+
+async function writeProfileUserId(
+  profile: string,
+  userId: string,
+): Promise<void> {
+  await mkdir(profileDir(profile), { recursive: true });
+  await writeFile(profileUserIdPath(profile), userId, "utf-8");
+}
+
+async function removeProfileUserId(profile: string): Promise<void> {
+  try {
+    await rm(profileUserIdPath(profile));
+  } catch {
+    // Already gone — nothing to do.
+  }
+}
+
+function normalizeUrl(u: string): string {
+  if (!u) return "";
+  try {
+    const parsed = new URL(u);
+    return `${parsed.protocol}//${parsed.host}`.toLowerCase();
+  } catch {
+    return u.replace(/\/+$/, "").toLowerCase();
+  }
+}
+
+function urlsMatch(a: string, b: string): boolean {
+  const na = normalizeUrl(a);
+  const nb = normalizeUrl(b);
+  return na.length > 0 && na === nb;
+}
+
+function sendStatus(status: DaemonStatus): void {
+  const win = getMainWindow();
+  win?.webContents.send("daemon:status", status);
+}
+
+interface HealthPayload {
+  status?: string;
+  pid?: number;
+  uptime?: string;
+  daemon_id?: string;
+  device_name?: string;
+  server_url?: string;
+  cli_version?: string;
+  active_task_count?: number;
+  agents?: string[];
+  workspaces?: unknown[];
+}
+
+async function fetchHealthAtPort(
+  port: number,
+): Promise<HealthPayload | null> {
+  try {
+    const controller = new AbortController();
+    const timeout = setTimeout(() => controller.abort(), 2_000);
+    const res = await fetch(`http://127.0.0.1:${port}/health`, {
+      signal: controller.signal,
+    });
+    clearTimeout(timeout);
+    if (!res.ok) return null;
+    return (await res.json()) as HealthPayload;
+  } catch {
+    return null;
+  }
+}
+
+// Desktop owns a dedicated CLI profile named after the target API host, so it
+// never reads or writes the user's hand-configured profiles. Profile dir:
+//   ~/.multica/profiles/desktop-<host>/
+function deriveProfileName(targetUrl: string): string {
+  try {
+    const url = new URL(targetUrl);
+    const host = url.host.replace(/:/g, "-").toLowerCase();
+    return `desktop-${host}`;
+  } catch {
+    return "desktop";
+  }
+}
+
+async function readProfileConfig(
+  profile: string,
+): Promise<Record<string, unknown>> {
+  try {
+    const raw = await readFile(profileConfigPath(profile), "utf-8");
+    const parsed = JSON.parse(raw);
+    return parsed && typeof parsed === "object" ? parsed : {};
+  } catch {
+    return {};
+  }
+}
+
+async function writeProfileConfig(
+  profile: string,
+  cfg: Record<string, unknown>,
+): Promise<void> {
+  const op = async () => {
+    await mkdir(profileDir(profile), { recursive: true });
+    await writeFile(
+      profileConfigPath(profile),
+      JSON.stringify(cfg, null, 2),
+      "utf-8",
+    );
+  };
+  const next = configWriteChain.catch(() => {}).then(op);
+  configWriteChain = next.catch(() => {});
+  return next;
+}
+
+/**
+ * Returns the Desktop-owned profile for the current target API URL. Creates
+ * the profile's config.json on demand with `server_url` pinned to the target.
+ *
+ * This function never falls back to the default profile, and never touches a
+ * profile whose name doesn't start with `desktop-`, so the user's manually
+ * configured CLI profiles are untouched.
+ */
+async function resolveActiveProfile(): Promise<ActiveProfile> {
+  const target = targetApiBaseUrl;
+  if (!target) return { name: "", port: DEFAULT_HEALTH_PORT };
+
+  const name = deriveProfileName(target);
+  const cfg = await readProfileConfig(name);
+
+  if (cfg.server_url !== target) {
+    cfg.server_url = target;
+    await writeProfileConfig(name, cfg);
+    console.log(`[daemon] initialized profile "${name}" → ${target}`);
+  }
+
+  return { name, port: healthPortForProfile(name) };
+}
+
+async function ensureActiveProfile(): Promise<ActiveProfile> {
+  if (activeProfile) return activeProfile;
+  activeProfile = await resolveActiveProfile();
+  return activeProfile;
+}
+
+function invalidateActiveProfile(): void {
+  activeProfile = null;
+}
+
+async function fetchHealth(): Promise<DaemonStatus> {
+  // While the CLI is being downloaded or has permanently failed, short-circuit
+  // polling — there's nothing to probe yet and /health calls would just return
+  // "stopped", which would overwrite the correct setup state in the UI.
+  if (currentState === "installing_cli" || currentState === "cli_not_found") {
+    return { state: currentState };
+  }
+
+  const active = await ensureActiveProfile();
+  const data = await fetchHealthAtPort(active.port);
+
+  if (!data || data.status !== "running") {
+    return {
+      state: currentState === "starting" ? "starting" : "stopped",
+      profile: active.name,
+    };
+  }
+
+  // Safety: if we have a target URL and the daemon on our port reports a
+  // different server_url, it's not "our" daemon — drop it and re-resolve.
+  if (
+    targetApiBaseUrl &&
+    data.server_url &&
+    !urlsMatch(data.server_url, targetApiBaseUrl)
+  ) {
+    invalidateActiveProfile();
+    return { state: "stopped" };
+  }
+
+  return {
+    state: "running",
+    pid: data.pid,
+    uptime: data.uptime,
+    daemonId: data.daemon_id,
+    deviceName: data.device_name,
+    agents: data.agents ?? [],
+    workspaceCount: Array.isArray(data.workspaces)
+      ? data.workspaces.length
+      : 0,
+    profile: active.name,
+    serverUrl: data.server_url,
+  };
+}
+
+function findCliOnPath(): string | null {
+  const candidates = process.platform === "win32" ? ["multica.exe"] : ["multica"];
+  const paths = (process.env["PATH"] ?? "").split(
+    process.platform === "win32" ? ";" : ":",
+  );
+  if (process.platform === "darwin") {
+    paths.push("/opt/homebrew/bin", "/usr/local/bin");
+  }
+  for (const name of candidates) {
+    for (const dir of paths) {
+      const full = join(dir, name);
+      if (existsSync(full)) return full;
+    }
+  }
+  return null;
+}
+
+/**
+ * Returns the path to the CLI binary bundled inside the Desktop app.
+ *
+ * - Dev (`electron-vite dev`): `app.getAppPath()` → `apps/desktop`, resolving
+ *   to `apps/desktop/resources/bin/multica`. `bundle-cli.mjs` populates this
+ *   before dev starts, so iterating on Go changes is "make build → restart".
+ * - Packaged: `app.getAppPath()` → `<Multica.app>/Contents/Resources/app.asar`.
+ *   electron-builder's `asarUnpack: resources/**` extracts the binary to
+ *   `app.asar.unpacked/`, so we swap the path segment to execute it.
+ */
+function bundledCliPath(): string {
+  const binName = process.platform === "win32" ? "multica.exe" : "multica";
+  return join(app.getAppPath(), "resources", "bin", binName).replace(
+    "app.asar",
+    "app.asar.unpacked",
+  );
+}
+
+async function probeCliBinary(
+  bin: string,
+  source: "bundled" | "managed" | "path",
+): Promise<string | null> {
+  try {
+    const stdout = await new Promise<string>((resolve, reject) => {
+      execFile(
+        bin,
+        ["version", "--output", "json"],
+        { timeout: 5_000 },
+        (err, out) => {
+          if (err) reject(err);
+          else resolve(out);
+        },
+      );
+    });
+    const parsed = JSON.parse(stdout) as { version?: string };
+    if (typeof parsed.version === "string" && parsed.version.length > 0) {
+      return parsed.version;
+    }
+    console.warn(
+      `[daemon] ignoring ${source} CLI at ${bin}: version output was missing or invalid`,
+    );
+    return null;
+  } catch (err) {
+    console.warn(`[daemon] ignoring ${source} CLI at ${bin}:`, err);
+    return null;
+  }
+}
+
+/**
+ * Returns a usable `multica` binary path. Priority:
+ *   1. Cached result from a previous successful resolve.
+ *   2. Bundled binary shipped with the Desktop app (`bundle-cli.mjs`).
+ *   3. Managed binary already installed in userData (`managedCliPath`).
+ *   4. Download + install latest release into userData.
+ *   5. `multica` on PATH (dev convenience / user-installed via brew).
+ * Returns `null` only when all of the above fail.
+ *
+ * Bundled is preferred so Desktop iterates in lockstep with Go changes in
+ * the same repo — avoids the 404 / stale-API problem when the Desktop's
+ * TS side is ahead of the last published CLI release.
+ *
+ * This function is idempotent and safe to call concurrently — in-flight
+ * installs are de-duplicated via `cliResolvePromise`.
+ */
+async function resolveCliBinary(): Promise<string | null> {
+  if (cachedCliBinary !== undefined) return cachedCliBinary;
+  if (cliResolvePromise) return cliResolvePromise;
+
+  cliResolvePromise = (async () => {
+    const bundled = bundledCliPath();
+    if (existsSync(bundled)) {
+      const version = await probeCliBinary(bundled, "bundled");
+      if (version) {
+        console.log(`[daemon] using bundled CLI at ${bundled}`);
+        cachedCliBinary = bundled;
+        cachedCliBinaryVersion = version;
+        return bundled;
+      }
+    }
+
+    const managed = managedCliPath();
+    if (existsSync(managed)) {
+      const version = await probeCliBinary(managed, "managed");
+      if (version) {
+        cachedCliBinary = managed;
+        cachedCliBinaryVersion = version;
+        return managed;
+      }
+    }
+
+    try {
+      const installed = await ensureManagedCli({
+        forceInstall: existsSync(managed),
+      });
+      const version = await probeCliBinary(installed, "managed");
+      if (version) {
+        cachedCliBinary = installed;
+        cachedCliBinaryVersion = version;
+        return installed;
+      }
+      console.warn(
+        `[daemon] managed CLI at ${installed} failed validation after install`,
+      );
+    } catch (err) {
+      console.warn("[daemon] CLI auto-install failed, falling back to PATH:", err);
+    }
+
+    const onPath = findCliOnPath();
+    if (onPath) {
+      const version = await probeCliBinary(onPath, "path");
+      if (version) {
+        cachedCliBinary = onPath;
+        cachedCliBinaryVersion = version;
+        return onPath;
+      }
+    }
+
+    cachedCliBinary = null;
+    cachedCliBinaryVersion = null;
+    return null;
+  })();
+
+  try {
+    return await cliResolvePromise;
+  } finally {
+    cliResolvePromise = null;
+  }
+}
+
+/**
+ * Reads the version of the currently resolved CLI binary. Cached for the
+ * process lifetime — the bundled binary doesn't change after bundle time.
+ * Returns null on any failure (unknown `go` at bundle time, broken binary,
+ * wrong-arch bundled binary, etc.) so callers can fail open.
+ */
+async function getCliBinaryVersion(): Promise<string | null> {
+  if (cachedCliBinaryVersion !== undefined) return cachedCliBinaryVersion;
+  const bin = await resolveCliBinary();
+  if (!bin) {
+    cachedCliBinaryVersion = null;
+    return null;
+  }
+  cachedCliBinaryVersion = await probeCliBinary(bin, "path");
+  return cachedCliBinaryVersion;
+}
+
+/**
+ * Compares the running daemon's `cli_version` against the CLI binary we
+ * would use to spawn a new one, and restarts only when safe. The decision
+ * logic itself is in `version-decision.ts` (pure, unit-tested); this
+ * wrapper handles the async plumbing and side effects.
+ *
+ * Restart is only fired when ALL of:
+ *   - a daemon is actually running on the active profile's port
+ *   - both sides report a version and the strings differ
+ *   - `active_task_count` is 0 (no in-flight agent work would be killed)
+ *
+ * On a confirmed mismatch while the daemon is busy, `pendingVersionRestart`
+ * is set; the poll loop retries this function on each 5s tick and will fire
+ * the restart as soon as the daemon drains.
+ */
+async function ensureRunningDaemonVersionMatches(): Promise<
+  "restarted" | "deferred" | "ok" | "not_running"
+> {
+  const active = await ensureActiveProfile();
+  const running = await fetchHealthAtPort(active.port);
+  const bundled = await getCliBinaryVersion();
+  const action = decideVersionAction(bundled, running);
+
+  switch (action) {
+    case "not_running":
+      pendingVersionRestart = false;
+      return "not_running";
+    case "ok":
+      pendingVersionRestart = false;
+      return "ok";
+    case "defer": {
+      if (!pendingVersionRestart) {
+        const activeTasks = running?.active_task_count ?? 0;
+        console.log(
+          `[daemon] CLI version mismatch (bundled=${bundled} running=${running?.cli_version}); deferring restart until ${activeTasks} active task(s) finish`,
+        );
+      }
+      pendingVersionRestart = true;
+      return "deferred";
+    }
+    case "restart":
+      console.log(
+        `[daemon] CLI version mismatch (bundled=${bundled} running=${running?.cli_version}) — restarting daemon`,
+      );
+      pendingVersionRestart = false;
+      await restartDaemon();
+      return "restarted";
+  }
+}
+
+/**
+ * Exchange the user's JWT for a long-lived PAT via POST /api/tokens. The
+ * daemon needs a PAT (or `mul_` / `mdt_` token) because JWTs expire in 30
+ * days and signatures are tied to a specific backend instance.
+ */
+async function mintPat(jwt: string): Promise<string> {
+  if (!targetApiBaseUrl) {
+    throw new Error("mint PAT: target API URL not set");
+  }
+  const url = `${targetApiBaseUrl.replace(/\/+$/, "")}/api/tokens`;
+  const res = await fetch(url, {
+    method: "POST",
+    headers: {
+      "Content-Type": "application/json",
+      Authorization: `Bearer ${jwt}`,
+    },
+    // Omit expires_in_days → server treats as null → non-expiring PAT.
+    body: JSON.stringify({ name: "Multica Desktop" }),
+  });
+  if (!res.ok) {
+    const body = await res.text().catch(() => "");
+    throw new Error(`mint PAT failed: ${res.status} ${res.statusText} ${body}`);
+  }
+  const data = (await res.json()) as { token?: unknown };
+  if (typeof data.token !== "string" || !data.token.startsWith("mul_")) {
+    throw new Error("mint PAT: response missing token");
+  }
+  return data.token;
+}
+
+/**
+ * Ensure the active profile's config.json has a usable token for the daemon.
+ *
+ * - Input from the renderer is the user's JWT (from localStorage) plus the
+ *   current user's id, so we can detect session changes.
+ * - If the profile already has a cached PAT (`mul_...`) AND the sidecar user
+ *   id matches the caller, reuse it — minting fresh on every launch would
+ *   accumulate garbage in the user's tokens page.
+ * - On user mismatch (or first run) call POST /api/tokens with the JWT to
+ *   mint a fresh PAT, overwriting any stale cached PAT. This is the critical
+ *   path: without it, a previous user's PAT would be used by a new session.
+ * - If the caller happens to pass a PAT directly, write it through.
+ * - When we mint fresh and a daemon is already running, restart it so the
+ *   new credentials take effect (the Go daemon reads config at startup).
+ */
+async function syncToken(
+  tokenFromRenderer: string,
+  userId: string,
+): Promise<void> {
+  const active = await ensureActiveProfile();
+  const config = await readProfileConfig(active.name);
+  const previousUserId = await readProfileUserId(active.name);
+  const userChanged = Boolean(previousUserId) && previousUserId !== userId;
+  const sameUserWithCachedPat =
+    !userChanged &&
+    previousUserId === userId &&
+    typeof config.token === "string" &&
+    config.token.startsWith("mul_");
+
+  let finalToken: string;
+  if (tokenFromRenderer.startsWith("mul_")) {
+    finalToken = tokenFromRenderer;
+  } else if (sameUserWithCachedPat) {
+    finalToken = config.token as string;
+  } else {
+    try {
+      finalToken = await mintPat(tokenFromRenderer);
+      console.log(
+        `[daemon] minted PAT for profile "${active.name}" (user_changed=${userChanged})`,
+      );
+    } catch (err) {
+      console.error("[daemon] failed to mint PAT:", err);
+      throw err;
+    }
+  }
+
+  config.token = finalToken;
+  if (targetApiBaseUrl) config.server_url = targetApiBaseUrl;
+  await writeProfileConfig(active.name, config);
+  await writeProfileUserId(active.name, userId);
+
+  // If we just rotated credentials onto a running daemon, restart it so the
+  // in-memory token in the Go process matches the new config.
+  if (userChanged) {
+    try {
+      const existing = await fetchHealthAtPort(active.port);
+      if (existing?.status === "running") {
+        console.log(
+          "[daemon] user switched — restarting daemon with new credentials",
+        );
+        void restartDaemon();
+      }
+    } catch (err) {
+      console.warn("[daemon] restart-on-user-switch failed:", err);
+    }
+  }
+}
+
+async function loadPrefs(): Promise<DaemonPrefs> {
+  try {
+    const raw = await readFile(PREFS_PATH, "utf-8");
+    const parsed = JSON.parse(raw);
+    return { ...DEFAULT_PREFS, ...parsed };
+  } catch {
+    return { ...DEFAULT_PREFS };
+  }
+}
+
+async function savePrefs(prefs: DaemonPrefs): Promise<void> {
+  const dir = join(homedir(), ".multica");
+  await mkdir(dir, { recursive: true });
+  await writeFile(PREFS_PATH, JSON.stringify(prefs, null, 2), "utf-8");
+}
+
+async function clearToken(): Promise<void> {
+  const active = await ensureActiveProfile();
+  const config = await readProfileConfig(active.name);
+  if ("token" in config) {
+    delete config.token;
+    await writeProfileConfig(active.name, config);
+  }
+  // Always drop the sidecar so a subsequent syncToken from any user is
+  // treated as a fresh mint, not a reuse of a stale cached PAT.
+  await removeProfileUserId(active.name);
+}
+
+async function withGuard<T>(fn: () => Promise<T>): Promise<T | { success: false; error: string }> {
+  if (operationInProgress) {
+    return { success: false, error: "Another daemon operation is in progress" };
+  }
+  operationInProgress = true;
+  try {
+    return await fn();
+  } finally {
+    operationInProgress = false;
+  }
+}
+
+function profileArgs(active: ActiveProfile): string[] {
+  return active.name ? ["--profile", active.name] : [];
+}
+
+// Env passed to every CLI child so the daemon process knows it was spawned
+// by the Desktop app. The server uses this to mark runtimes as managed and
+// hide CLI self-update UI. Computed lazily so it picks up the PATH fix
+// applied by fix-path in main/index.ts — as a top-level const it would
+// snapshot process.env at import time, before that block runs.
+function desktopSpawnEnv(): NodeJS.ProcessEnv {
+  return { ...process.env, MULTICA_LAUNCHED_BY: "desktop" };
+}
+
+async function startDaemon(): Promise<{ success: boolean; error?: string }> {
+  const bin = await resolveCliBinary();
+  if (!bin) return { success: false, error: "multica CLI is not installed" };
+
+  const active = await ensureActiveProfile();
+  const existing = await fetchHealthAtPort(active.port);
+  if (existing?.status === "running") {
+    pollOnce();
+    return { success: true };
+  }
+
+  currentState = "starting";
+  sendStatus({ state: "starting" });
+
+  const args = ["daemon", "start", ...profileArgs(active)];
+
+  return new Promise((resolve) => {
+    execFile(
+      bin,
+      args,
+      { timeout: 20_000, env: desktopSpawnEnv() },
+      (err) => {
+        if (err) {
+          currentState = "stopped";
+          sendStatus({ state: "stopped" });
+          resolve({ success: false, error: err.message });
+          return;
+        }
+        // Stay in "starting" until pollOnce confirms /health — the CLI
+        // returning 0 only means the supervisor was spawned, not that the
+        // daemon process is already listening.
+        pollOnce();
+        resolve({ success: true });
+      },
+    );
+  });
+}
+
+async function stopDaemon(): Promise<{ success: boolean; error?: string }> {
+  const bin = await resolveCliBinary();
+  if (!bin) return { success: false, error: "multica CLI is not installed" };
+
+  const active = await ensureActiveProfile();
+  currentState = "stopping";
+  sendStatus({ state: "stopping" });
+
+  const args = ["daemon", "stop", ...profileArgs(active)];
+
+  return new Promise((resolve) => {
+    execFile(bin, args, { timeout: 15_000 }, (err) => {
+      if (err) {
+        resolve({ success: false, error: err.message });
+      } else {
+        resolve({ success: true });
+      }
+      currentState = "stopped";
+      sendStatus({ state: "stopped" });
+    });
+  });
+}
+
+async function restartDaemon(): Promise<{ success: boolean; error?: string }> {
+  const stopResult = await stopDaemon();
+  if (!stopResult.success) return stopResult;
+  return startDaemon();
+}
+
+async function pollOnce(): Promise<void> {
+  const status = await fetchHealth();
+  currentState = status.state;
+  sendStatus(status);
+  // Retry a deferred version-mismatch restart once the daemon drains.
+  if (pendingVersionRestart && status.state === "running") {
+    void ensureRunningDaemonVersionMatches();
+  }
+}
+
+function startPolling(): void {
+  if (statusPollTimer) return;
+  pollOnce();
+  statusPollTimer = setInterval(pollOnce, POLL_INTERVAL_MS);
+}
+
+/**
+ * Ensures the CLI binary is available, then transitions into the normal
+ * stopped/running state machine. Called once at startup and again on
+ * user-triggered `daemon:retry-install`.
+ */
+async function bootstrapCli(): Promise<void> {
+  const bin = await resolveCliBinary();
+  if (!bin) {
+    currentState = "cli_not_found";
+    sendStatus({ state: "cli_not_found" });
+    return;
+  }
+  currentState = "stopped";
+  sendStatus({ state: "stopped" });
+  startPolling();
+}
+
+function stopPolling(): void {
+  if (statusPollTimer) {
+    clearInterval(statusPollTimer);
+    statusPollTimer = null;
+  }
+}
+
+const LOG_TAIL_INITIAL_WINDOW_BYTES = 32 * 1024;
+const LOG_TAIL_INITIAL_LINES = 200;
+const LOG_TAIL_POLL_MS = 500;
+
+async function readLogRange(
+  path: string,
+  startAt: number,
+  length: number,
+): Promise<string> {
+  const handle = await open(path, "r");
+  try {
+    const buffer = Buffer.alloc(length);
+    const { bytesRead } = await handle.read(buffer, 0, length, startAt);
+    return buffer.subarray(0, bytesRead).toString("utf-8");
+  } finally {
+    await handle.close();
+  }
+}
+
+function sendLines(win: BrowserWindow, text: string): void {
+  const lines = text.split("\n").filter((line) => line.length > 0);
+  for (const line of lines) {
+    win.webContents.send("daemon:log-line", line);
+  }
+}
+
+// Cross-platform tail -f replacement: read the tail of the file once, then
+// poll its stat with fs.watchFile and forward any new bytes since the last
+// known offset. watchFile works on macOS, Linux, and Windows; spawn("tail")
+// would silently fail on Windows.
+function startLogTail(win: BrowserWindow, retryCount = 0): void {
+  stopLogTail();
+
+  void ensureActiveProfile().then(async (active) => {
+    const logPath = profileLogPath(active.name);
+    if (!existsSync(logPath)) {
+      if (retryCount < LOG_TAIL_MAX_RETRIES) {
+        setTimeout(() => startLogTail(win, retryCount + 1), LOG_TAIL_RETRY_MS);
+      }
+      return;
+    }
+
+    let position = 0;
+    try {
+      const initialStats = await stat(logPath);
+      const windowBytes = Math.min(
+        initialStats.size,
+        LOG_TAIL_INITIAL_WINDOW_BYTES,
+      );
+      const startAt = initialStats.size - windowBytes;
+      if (windowBytes > 0) {
+        const text = await readLogRange(logPath, startAt, windowBytes);
+        const lines = text
+          .split("\n")
+          .filter((line) => line.length > 0)
+          .slice(-LOG_TAIL_INITIAL_LINES);
+        for (const line of lines) {
+          win.webContents.send("daemon:log-line", line);
+        }
+      }
+      position = initialStats.size;
+    } catch (err) {
+      console.warn("[daemon] log tail initial read failed:", err);
+      return;
+    }
+
+    const listener: StatsListener = (curr) => {
+      const target = getMainWindow();
+      if (!target) return;
+      // File rotated/truncated — restart from the new beginning.
+      if (curr.size < position) position = 0;
+      if (curr.size === position) return;
+      const from = position;
+      const length = curr.size - from;
+      position = curr.size;
+      readLogRange(logPath, from, length)
+        .then((text) => sendLines(target, text))
+        .catch((err) => {
+          console.warn("[daemon] log tail read failed:", err);
+        });
+    };
+
+    watchFile(logPath, { interval: LOG_TAIL_POLL_MS }, listener);
+    logTailWatcher = { path: logPath, listener };
+  });
+}
+
+function stopLogTail(): void {
+  if (logTailWatcher) {
+    unwatchFile(logTailWatcher.path, logTailWatcher.listener);
+    logTailWatcher = null;
+  }
+}
+
+export function setupDaemonManager(
+  windowGetter: () => BrowserWindow | null,
+): void {
+  getMainWindow = windowGetter;
+
+  ipcMain.handle("daemon:set-target-api-url", async (_e, url: string) => {
+    const normalized = url || null;
+    if (targetApiBaseUrl !== normalized) {
+      console.log(`[daemon] target API URL set to ${normalized ?? "(none)"}`);
+      targetApiBaseUrl = normalized;
+      invalidateActiveProfile();
+      await pollOnce();
+    }
+  });
+  ipcMain.handle("daemon:start", () => withGuard(() => startDaemon()));
+  ipcMain.handle("daemon:stop", () => withGuard(() => stopDaemon()));
+  ipcMain.handle("daemon:restart", () => withGuard(() => restartDaemon()));
+  ipcMain.handle("daemon:get-status", () => fetchHealth());
+  ipcMain.handle(
+    "daemon:sync-token",
+    (_event, token: string, userId: string) => syncToken(token, userId),
+  );
+  ipcMain.handle("daemon:clear-token", () => clearToken());
+  ipcMain.handle("daemon:is-cli-installed", async () => {
+    const bin = await resolveCliBinary();
+    return bin !== null;
+  });
+  ipcMain.handle("daemon:retry-install", async () => {
+    cachedCliBinary = undefined;
+    cliResolvePromise = null;
+    // A retry-install may land a new CLI at a different version; drop the
+    // cached version string so the next check re-reads the binary.
+    cachedCliBinaryVersion = undefined;
+    await bootstrapCli();
+  });
+  ipcMain.handle("daemon:get-prefs", () => loadPrefs());
+  ipcMain.handle(
+    "daemon:set-prefs",
+    (_event, prefs: Partial<DaemonPrefs>) =>
+      loadPrefs().then((cur) => {
+        const merged = { ...cur, ...prefs };
+        return savePrefs(merged).then(() => merged);
+      }),
+  );
+  ipcMain.handle("daemon:auto-start", async () => {
+    const prefs = await loadPrefs();
+    if (!prefs.autoStart) return;
+    const bin = await resolveCliBinary();
+    if (!bin) return;
+    const health = await fetchHealth();
+    if (health.state === "running") {
+      // Daemon is up but may be running an older CLI than the one we just
+      // bundled. Restart it so the new binary actually takes effect.
+      await ensureRunningDaemonVersionMatches();
+      return;
+    }
+    await startDaemon();
+  });
+
+  ipcMain.on("daemon:start-log-stream", () => {
+    const win = getMainWindow();
+    if (win) startLogTail(win);
+  });
+
+  ipcMain.on("daemon:stop-log-stream", () => {
+    stopLogTail();
+  });
+
+  // Reveal the daemon's log file in the user's default editor / Console
+  // app. Acts as the escape hatch when the in-app log viewer isn't enough
+  // (full history, complex search, copy-to-clipboard at scale).
+  ipcMain.handle("daemon:open-log-file", async () => {
+    const active = await ensureActiveProfile();
+    const logPath = profileLogPath(active.name);
+    if (!existsSync(logPath)) {
+      return { success: false, error: "Log file not found yet" };
+    }
+    // shell.openPath returns "" on success, error string on failure.
+    const error = await shell.openPath(logPath);
+    return error === "" ? { success: true } : { success: false, error };
+  });
+
+  // First-run CLI install kicks off here. Status bar shows "Setting up…"
+  // until the managed binary is on disk (instant on subsequent launches).
+  currentState = "installing_cli";
+  sendStatus({ state: "installing_cli" });
+  void bootstrapCli();
+
+  let isQuitting = false;
+  app.on("before-quit", (event) => {
+    if (isQuitting) return;
+    stopPolling();
+    stopLogTail();
+
+    loadPrefs().then(async (prefs) => {
+      if (prefs.autoStop) {
+        isQuitting = true;
+        event.preventDefault();
+        try {
+          await stopDaemon();
+        } catch {
+          // Best-effort stop on quit
+        }
+        app.quit();
+      }
+    });
+  });
+}

apps/desktop/src/main/external-url.test.ts

@@ -0,0 +1,73 @@
+import { describe, expect, it, vi, beforeEach } from "vitest";
+
+vi.mock("electron", () => ({
+  shell: { openExternal: vi.fn().mockResolvedValue(undefined) },
+}));
+
+import { shell } from "electron";
+import { isSafeExternalHttpUrl, openExternalSafely } from "./external-url";
+
+describe("isSafeExternalHttpUrl", () => {
+  it("allows http and https URLs", () => {
+    expect(isSafeExternalHttpUrl("https://multica.ai")).toBe(true);
+    expect(isSafeExternalHttpUrl("http://localhost:3000/auth")).toBe(true);
+  });
+
+  it("allows https URLs with embedded credentials", () => {
+    // WHATWG URL parses these as https; OS-level handling is the shell's concern.
+    expect(isSafeExternalHttpUrl("https://user:pass@example.com")).toBe(true);
+  });
+
+  it("normalizes scheme casing so uppercase variants can't bypass", () => {
+    expect(isSafeExternalHttpUrl("HTTPS://example.com")).toBe(true);
+    expect(isSafeExternalHttpUrl("FILE:///etc/passwd")).toBe(false);
+  });
+
+  it("rejects dangerous pseudo-schemes", () => {
+    expect(isSafeExternalHttpUrl("javascript:alert(1)")).toBe(false);
+    expect(
+      isSafeExternalHttpUrl("data:text/html,<script>alert(1)</script>"),
+    ).toBe(false);
+  });
+
+  it("rejects filesystem and network transport schemes", () => {
+    expect(isSafeExternalHttpUrl("file:///etc/passwd")).toBe(false);
+    expect(isSafeExternalHttpUrl("ftp://example.com/x")).toBe(false);
+    expect(isSafeExternalHttpUrl("smb://share/x")).toBe(false);
+  });
+
+  it("rejects local-handler schemes used in past RCE chains", () => {
+    expect(isSafeExternalHttpUrl("vscode://file/test")).toBe(false);
+    expect(isSafeExternalHttpUrl("ms-msdt:/id%20PCWDiagnostic")).toBe(false);
+  });
+
+  it("rejects mailto and other non-web schemes", () => {
+    expect(isSafeExternalHttpUrl("mailto:test@example.com")).toBe(false);
+    expect(isSafeExternalHttpUrl("tel:+15551234567")).toBe(false);
+  });
+
+  it("rejects empty, whitespace, and malformed input", () => {
+    expect(isSafeExternalHttpUrl("")).toBe(false);
+    expect(isSafeExternalHttpUrl(" ")).toBe(false);
+    expect(isSafeExternalHttpUrl("not a url")).toBe(false);
+    expect(isSafeExternalHttpUrl("http://")).toBe(false);
+  });
+});
+
+describe("openExternalSafely", () => {
+  beforeEach(() => {
+    vi.mocked(shell.openExternal).mockClear();
+  });
+
+  it("forwards http/https URLs to shell.openExternal", () => {
+    openExternalSafely("https://multica.ai");
+    expect(shell.openExternal).toHaveBeenCalledWith("https://multica.ai");
+  });
+
+  it("does not call shell.openExternal for rejected schemes", () => {
+    openExternalSafely("file:///etc/passwd");
+    openExternalSafely("javascript:alert(1)");
+    openExternalSafely("not a url");
+    expect(shell.openExternal).not.toHaveBeenCalled();
+  });
+});

apps/desktop/src/main/external-url.ts

@@ -0,0 +1,38 @@
+import { shell } from "electron";
+
+// True when the URL parses and uses http/https — the only schemes we let
+// reach `shell.openExternal`. Scheme comparison is safe because the WHATWG
+// URL parser lowercases the protocol field.
+export function isSafeExternalHttpUrl(url: string): boolean {
+  return getHttpProtocol(url) !== null;
+}
+
+// Canonical wrapper around shell.openExternal. All renderer-controlled URLs
+// that eventually reach the OS shell MUST flow through here; direct calls
+// to `shell.openExternal` elsewhere in the main process are banned by the
+// no-restricted-syntax rule in apps/desktop/eslint.config.mjs.
+export function openExternalSafely(url: string): Promise<void> | void {
+  if (getHttpProtocol(url) === null) {
+    console.warn(`[security] blocked openExternal: ${describeScheme(url)}`);
+    return;
+  }
+  return shell.openExternal(url);
+}
+
+function getHttpProtocol(url: string): "http:" | "https:" | null {
+  try {
+    const { protocol } = new URL(url);
+    if (protocol === "http:" || protocol === "https:") return protocol;
+    return null;
+  } catch {
+    return null;
+  }
+}
+
+function describeScheme(url: string): string {
+  try {
+    return `scheme=${new URL(url).protocol}`;
+  } catch {
+    return "invalid URL";
+  }
+}

apps/desktop/src/main/index.ts

@@ -0,0 +1,384 @@
+import { app, BrowserWindow, ipcMain, nativeImage, Notification } from "electron";
+import { homedir } from "os";
+import { join } from "path";
+import { electronApp, optimizer, is } from "@electron-toolkit/utils";
+import fixPath from "fix-path";
+import { setupAutoUpdater } from "./updater";
+import { setupDaemonManager } from "./daemon-manager";
+import { openExternalSafely } from "./external-url";
+import { installContextMenu } from "./context-menu";
+import { getAppVersion } from "./app-version";
+import { loadRuntimeConfig } from "./runtime-config-loader";
+import type { RuntimeConfigResult } from "../shared/runtime-config";
+
+// Bundled icon used for dev-mode dock/taskbar branding. In production the
+// app bundle icon (from electron-builder) wins; this path is only consumed
+// by the `is.dev` branch below.
+const DEV_ICON_PATH = join(__dirname, "../../resources/icon.png");
+
+// macOS/Linux GUI launches inherit a minimal PATH from launchd that omits
+// the user's shell config (~/.zshrc, Homebrew, nvm, ~/.local/bin, etc.).
+// Run the user's login shell once to recover the real PATH so the bundled
+// multica CLI can find agent binaries like claude/codex/opencode. Must run
+// before any child_process.spawn / execFile call in the main process —
+// ES module imports are hoisted, so this block executes before createWindow
+// or any daemon-manager spawn.
+if (process.platform !== "win32") {
+  fixPath();
+  // Fallback: prepend common install locations in case fix-path came up
+  // short (broken shell rc, non-interactive $SHELL, missing entries). Safe
+  // to duplicate — PATH lookups short-circuit on first match.
+  const fallbackPaths = [
+    "/opt/homebrew/bin",
+    "/usr/local/bin",
+    join(homedir(), ".local/bin"),
+  ];
+  process.env.PATH = `${fallbackPaths.join(":")}:${process.env.PATH ?? ""}`;
+}
+
+const PROTOCOL = "multica";
+
+let mainWindow: BrowserWindow | null = null;
+let runtimeConfigResult: RuntimeConfigResult = {
+  ok: false,
+  error: { message: "Runtime config has not loaded yet" },
+};
+
+// --- Deep link helpers ---------------------------------------------------
+
+function handleDeepLink(url: string): void {
+  try {
+    const parsed = new URL(url);
+    if (parsed.protocol !== `${PROTOCOL}:`) return;
+
+    // multica://auth/callback?token=<jwt>
+    if (parsed.hostname === "auth" && parsed.pathname === "/callback") {
+      const token = parsed.searchParams.get("token");
+      if (token && mainWindow) {
+        mainWindow.webContents.send("auth:token", token);
+      }
+      return;
+    }
+
+    // multica://invite/<invitationId>
+    // Dispatched from the web invite page when the user chooses "Open in
+    // desktop app". The renderer opens the invite overlay — no tab, no
+    // route persistence, so deep-linking the same invite twice stays safe.
+    if (parsed.hostname === "invite") {
+      const id = parsed.pathname.replace(/^\//, "");
+      if (id && mainWindow) {
+        mainWindow.webContents.send("invite:open", decodeURIComponent(id));
+      }
+      return;
+    }
+  } catch {
+    // Ignore malformed URLs
+  }
+}
+
+// --- Window creation -----------------------------------------------------
+
+// Tracks the OS-preferred language as last seen by the running process.
+// Updated on each window-focus check so we can emit a `locale:system-changed`
+// event to the renderer when the user changes their OS language without
+// quitting the app — without restart, app.getPreferredSystemLanguages()
+// would still report the boot value forever.
+let lastKnownSystemLocale = "en";
+
+function getSystemLocale(): string {
+  return app.getPreferredSystemLanguages()[0] ?? "en";
+}
+
+function createWindow(): void {
+  // Pass the OS-preferred language to the renderer via additionalArguments
+  // instead of a sync IPC call. process.argv is available to the preload
+  // script before the first network request, so the renderer's i18next
+  // instance can initialize with the right locale on the very first paint.
+  const systemLocale = getSystemLocale();
+  lastKnownSystemLocale = systemLocale;
+
+  mainWindow = new BrowserWindow({
+    width: 1280,
+    height: 800,
+    minWidth: 900,
+    minHeight: 600,
+    titleBarStyle: "hiddenInset",
+    trafficLightPosition: { x: 16, y: 13 },
+    show: false,
+    autoHideMenuBar: true,
+    // Windows/Linux pick up the window/taskbar icon from this option in
+    // dev — on macOS it's ignored (dock comes from app.dock.setIcon below).
+    ...(is.dev ? { icon: DEV_ICON_PATH } : {}),
+    webPreferences: {
+      preload: join(__dirname, "../preload/index.js"),
+      sandbox: false,
+      webSecurity: false,
+      additionalArguments: [`--multica-locale=${systemLocale}`],
+    },
+  });
+
+  // Strip Origin header from WebSocket upgrade requests so the server's
+  // origin whitelist doesn't reject connections from localhost dev origins.
+  mainWindow.webContents.session.webRequest.onBeforeSendHeaders(
+    { urls: ["wss://*/*", "ws://*/*"] },
+    (details, callback) => {
+      delete details.requestHeaders["Origin"];
+      callback({ requestHeaders: details.requestHeaders });
+    },
+  );
+
+  mainWindow.on("ready-to-show", () => {
+    mainWindow?.show();
+  });
+
+  // Detect OS language changes while the app is running. Electron has no
+  // dedicated event for this on any platform, so we poll on focus regain —
+  // catches the common case where users switch System Settings → Language
+  // and bring the app back. The renderer decides whether to act (it ignores
+  // the signal when the user has an explicit Settings choice).
+  mainWindow.on("focus", () => {
+    const current = getSystemLocale();
+    if (current === lastKnownSystemLocale) return;
+    lastKnownSystemLocale = current;
+    mainWindow?.webContents.send("locale:system-changed", current);
+  });
+
+  mainWindow.webContents.setWindowOpenHandler((details) => {
+    openExternalSafely(details.url);
+    return { action: "deny" };
+  });
+
+  // Prevent Cmd+R / Ctrl+R / Shift+Cmd+R / Shift+Ctrl+R / F5 from
+  // reloading the page. In a desktop app an accidental reload destroys
+  // in-memory state (tabs, drafts, WS connections) with no URL bar to
+  // navigate back. DevTools refresh (via the DevTools UI) still works.
+  mainWindow.webContents.on("before-input-event", (_event, input) => {
+    if (input.type !== "keyDown") return;
+    const cmdOrCtrl =
+      process.platform === "darwin" ? input.meta : input.control;
+    if (
+      (cmdOrCtrl && input.key.toLowerCase() === "r") ||
+      input.key === "F5"
+    ) {
+      _event.preventDefault();
+    }
+  });
+
+  installContextMenu(mainWindow.webContents);
+
+  if (is.dev && process.env["ELECTRON_RENDERER_URL"]) {
+    mainWindow.loadURL(process.env["ELECTRON_RENDERER_URL"]);
+  } else {
+    mainWindow.loadFile(join(__dirname, "../renderer/index.html"));
+  }
+}
+
+// --- Dev / production isolation -------------------------------------------
+// Give dev mode a separate app name and userData path so it gets its own
+// single-instance lock file and doesn't conflict with the packaged production
+// app. Must run BEFORE requestSingleInstanceLock() because the lock location
+// is derived from the userData path. (Same approach VS Code uses for
+// Stable / Insiders coexistence.)
+
+// DESKTOP_APP_SUFFIX lets parallel worktrees run dev Electron side-by-side
+// without fighting for the shared single-instance lock. The suffix is
+// appended to the app name + userData path, so each worktree gets its own
+// lock file. Default (no env var) keeps behavior unchanged — the common
+// single-worktree case still lands at "Multica Canary".
+const DEV_APP_NAME = process.env.DESKTOP_APP_SUFFIX
+  ? `Multica Canary ${process.env.DESKTOP_APP_SUFFIX}`
+  : "Multica Canary";
+
+if (is.dev) {
+  app.setName(DEV_APP_NAME);
+  app.setPath("userData", join(app.getPath("appData"), DEV_APP_NAME));
+}
+
+// --- Protocol registration -----------------------------------------------
+
+if (process.defaultApp) {
+  // In dev, register with the path to the electron binary + app path
+  app.setAsDefaultProtocolClient(PROTOCOL, process.execPath, [
+    app.getAppPath(),
+  ]);
+} else {
+  app.setAsDefaultProtocolClient(PROTOCOL);
+}
+
+// --- Single instance lock ------------------------------------------------
+
+const gotTheLock = app.requestSingleInstanceLock();
+
+if (!gotTheLock) {
+  app.quit();
+} else {
+  // Windows/Linux: second instance passes deep link via argv
+  app.on("second-instance", (_event, argv) => {
+    if (mainWindow) {
+      if (mainWindow.isMinimized()) mainWindow.restore();
+      mainWindow.focus();
+    }
+
+    // On Windows the deep link URL is the last argv entry
+    const deepLinkUrl = argv.find((arg) => arg.startsWith(`${PROTOCOL}://`));
+    if (deepLinkUrl) handleDeepLink(deepLinkUrl);
+  });
+
+  app.whenReady().then(async () => {
+    const viteEnv = import.meta.env as ImportMetaEnv & {
+      readonly VITE_API_URL?: string;
+      readonly VITE_WS_URL?: string;
+      readonly VITE_APP_URL?: string;
+    };
+
+    runtimeConfigResult = await loadRuntimeConfig({
+      isDev: is.dev,
+      // electron-vite exposes VITE_* on import.meta.env for the main process;
+      // keep dev URL overrides on the same source the renderer used before
+      // runtime config moved endpoint resolution into main/preload.
+      env: {
+        apiUrl: viteEnv.VITE_API_URL,
+        wsUrl: viteEnv.VITE_WS_URL,
+        appUrl: viteEnv.VITE_APP_URL,
+      },
+    });
+
+    electronApp.setAppUserModelId(
+      is.dev ? "ai.multica.desktop.dev" : "ai.multica.desktop",
+    );
+
+    // macOS: replace the default Electron dock icon with the bundled logo
+    // so the Canary dev build is visually distinct from a stock Electron
+    // run. `app.dock` is macOS-only — guard the call.
+    if (is.dev && process.platform === "darwin" && app.dock) {
+      const icon = nativeImage.createFromPath(DEV_ICON_PATH);
+      if (!icon.isEmpty()) app.dock.setIcon(icon);
+    }
+
+    app.on("browser-window-created", (_, window) => {
+      optimizer.watchWindowShortcuts(window);
+    });
+
+    // IPC: open URL in default browser (used by renderer for Google login).
+    // All scheme-allowlist enforcement lives in openExternalSafely — this
+    // is the single audit point for renderer-controlled URLs reaching the
+    // OS shell under the app's intentional webSecurity: false + sandbox:
+    // false configuration.
+    ipcMain.handle("shell:openExternal", (_event, url: string) => {
+      return openExternalSafely(url);
+    });
+
+    // Sync IPC: app version + normalized OS for preload. Sync (not invoke) so
+    // preload can attach the values to `desktopAPI.appInfo` before any renderer
+    // code reads them, ensuring the very first HTTP request from the renderer
+    // already carries X-Client-Version and X-Client-OS.
+    ipcMain.on("app:get-info", (event) => {
+      const p = process.platform;
+      const os = p === "darwin" ? "macos" : p === "win32" ? "windows" : p === "linux" ? "linux" : "unknown";
+      event.returnValue = { version: getAppVersion(), os };
+    });
+
+    // Sync IPC: preload exposes the validated runtime config before renderer
+    // boot. If desktop.json exists but is invalid, renderer receives the
+    // blocking error and must not silently fall back to the cloud defaults.
+    ipcMain.on("runtime-config:get", (event) => {
+      event.returnValue = runtimeConfigResult;
+    });
+
+    // IPC: toggle immersive mode — hides the macOS traffic lights so full-screen
+    // modals (e.g. create-workspace) can place UI in the top-left corner
+    // without fighting the native window controls' hit-test.
+    ipcMain.handle("window:setImmersive", (_event, immersive: boolean) => {
+      if (process.platform !== "darwin") return;
+      mainWindow?.setWindowButtonVisibility(!immersive);
+    });
+
+    // IPC: show a native OS notification for a new inbox item. The renderer
+    // only fires this when the app is unfocused (it gates on
+    // `document.hasFocus()`), so we don't fight macOS foreground suppression
+    // here. Clicking the banner focuses the main window and routes to the
+    // inbox item via a renderer-side listener.
+    ipcMain.on(
+      "notification:show",
+      (
+        _event,
+        {
+          slug,
+          itemId,
+          issueKey,
+          title,
+          body,
+        }: {
+          slug: string;
+          itemId: string;
+          issueKey: string;
+          title: string;
+          body: string;
+        },
+      ) => {
+        if (!Notification.isSupported()) return;
+        const notification = new Notification({ title, body });
+        notification.on("click", () => {
+          if (!mainWindow) return;
+          if (mainWindow.isMinimized()) mainWindow.restore();
+          mainWindow.show();
+          mainWindow.focus();
+          // Ship the full context back — the renderer pins the route to the
+          // source workspace (slug), marks the row read (itemId), and uses
+          // issueKey as the ?issue=<…> selector.
+          mainWindow.webContents.send("inbox:open", {
+            slug,
+            itemId,
+            issueKey,
+          });
+        });
+        notification.show();
+      },
+    );
+
+    // IPC: update the dock / taskbar unread badge. Values above 99 render as
+    // "99+". macOS is the primary target (user-visible dock badge); Linux
+    // Unity launchers also respect `setBadgeCount`. Windows' taskbar overlay
+    // needs a pre-rendered PNG and is deferred — the OS notification + the
+    // in-app inbox sidebar cover the core UX there for now.
+    ipcMain.on("badge:set", (_event, rawCount: number) => {
+      const count = Math.max(0, Math.floor(rawCount));
+      if (process.platform === "darwin") {
+        const label = count === 0 ? "" : count > 99 ? "99+" : String(count);
+        app.dock?.setBadge(label);
+      } else {
+        app.setBadgeCount(count);
+      }
+    });
+
+    createWindow();
+
+    setupAutoUpdater(() => mainWindow);
+    setupDaemonManager(() => mainWindow);
+
+    // macOS: deep link arrives via open-url event
+    app.on("open-url", (_event, url) => {
+      if (mainWindow) {
+        if (mainWindow.isMinimized()) mainWindow.restore();
+        mainWindow.focus();
+      }
+      handleDeepLink(url);
+    });
+
+    app.on("activate", () => {
+      if (BrowserWindow.getAllWindows().length === 0) createWindow();
+    });
+  });
+
+  // Check argv for deep link on cold start (Windows/Linux)
+  const deepLinkArg = process.argv.find((arg) =>
+    arg.startsWith(`${PROTOCOL}://`),
+  );
+  if (deepLinkArg) {
+    app.whenReady().then(() => handleDeepLink(deepLinkArg));
+  }
+}
+
+app.on("window-all-closed", () => {
+  if (process.platform !== "darwin") app.quit();
+});

apps/desktop/src/main/runtime-config-loader.test.ts

@@ -0,0 +1,90 @@
+import { mkdtemp, writeFile } from "fs/promises";
+import { join } from "path";
+import { tmpdir } from "os";
+import { describe, expect, it } from "vitest";
+import { loadRuntimeConfig } from "./runtime-config-loader";
+
+describe("loadRuntimeConfig", () => {
+  it("uses dev env and ignores desktop.json during electron-vite dev", async () => {
+    const dir = await mkdtemp(join(tmpdir(), "multica-desktop-config-"));
+    const configPath = join(dir, "desktop.json");
+    await writeFile(
+      configPath,
+      JSON.stringify({ schemaVersion: 1, apiUrl: "https://prod.example.com" }),
+    );
+
+    await expect(
+      loadRuntimeConfig({
+        isDev: true,
+        configPath,
+        env: {
+          apiUrl: "http://localhost:8080",
+          wsUrl: "ws://localhost:8080/ws",
+          appUrl: "http://localhost:3000",
+        },
+      }),
+    ).resolves.toEqual({
+      ok: true,
+      config: {
+        schemaVersion: 1,
+        apiUrl: "http://localhost:8080",
+        wsUrl: "ws://localhost:8080/ws",
+        appUrl: "http://localhost:3000",
+      },
+    });
+  });
+
+  it("uses cloud defaults when packaged config is absent", async () => {
+    const dir = await mkdtemp(join(tmpdir(), "multica-desktop-config-"));
+    await expect(
+      loadRuntimeConfig({
+        isDev: false,
+        configPath: join(dir, "missing.json"),
+        env: {},
+      }),
+    ).resolves.toEqual({
+      ok: true,
+      config: {
+        schemaVersion: 1,
+        apiUrl: "https://api.multica.ai",
+        wsUrl: "wss://api.multica.ai/ws",
+        appUrl: "https://multica.ai",
+      },
+    });
+  });
+
+  it("parses a valid packaged desktop.json", async () => {
+    const dir = await mkdtemp(join(tmpdir(), "multica-desktop-config-"));
+    const configPath = join(dir, "desktop.json");
+    await writeFile(
+      configPath,
+      JSON.stringify({ schemaVersion: 1, apiUrl: "https://api.example.com" }),
+    );
+
+    await expect(
+      loadRuntimeConfig({ isDev: false, configPath, env: {} }),
+    ).resolves.toEqual({
+      ok: true,
+      config: {
+        schemaVersion: 1,
+        apiUrl: "https://api.example.com",
+        wsUrl: "wss://api.example.com/ws",
+        appUrl: "https://example.com",
+      },
+    });
+  });
+
+  it("fails closed when packaged desktop.json is invalid", async () => {
+    const dir = await mkdtemp(join(tmpdir(), "multica-desktop-config-"));
+    const configPath = join(dir, "desktop.json");
+    await writeFile(configPath, "{");
+
+    const result = await loadRuntimeConfig({ isDev: false, configPath, env: {} });
+
+    expect(result.ok).toBe(false);
+    if (!result.ok) {
+      expect(result.error.message).toContain(configPath);
+      expect(result.error.message).toContain("Invalid desktop runtime config JSON");
+    }
+  });
+});

apps/desktop/src/main/runtime-config-loader.ts

@@ -0,0 +1,60 @@
+import { app } from "electron";
+import { readFile } from "fs/promises";
+import { join } from "path";
+import {
+  DEFAULT_RUNTIME_CONFIG,
+  parseRuntimeConfig,
+  runtimeConfigFromDevEnv,
+  type RuntimeConfig,
+  type RuntimeConfigEnv,
+  type RuntimeConfigResult,
+} from "../shared/runtime-config";
+
+export async function loadRuntimeConfig(options: {
+  isDev: boolean;
+  env: RuntimeConfigEnv;
+  configPath?: string;
+}): Promise<RuntimeConfigResult> {
+  if (options.isDev) {
+    try {
+      return { ok: true, config: runtimeConfigFromDevEnv(options.env) };
+    } catch (err) {
+      return { ok: false, error: { message: errorMessage(err) } };
+    }
+  }
+
+  const configPath = options.configPath ?? desktopConfigPath();
+  try {
+    const raw = await readFile(configPath, "utf-8");
+    return { ok: true, config: parseRuntimeConfig(raw) };
+  } catch (err) {
+    if (isMissingFileError(err)) {
+      return { ok: true, config: { ...DEFAULT_RUNTIME_CONFIG } };
+    }
+    return {
+      ok: false,
+      error: {
+        message: `Invalid ${configPath}: ${errorMessage(err)}`,
+      },
+    };
+  }
+}
+
+export function desktopConfigPath(): string {
+  return join(app.getPath("home"), ".multica", "desktop.json");
+}
+
+function isMissingFileError(err: unknown): boolean {
+  return Boolean(
+    err &&
+      typeof err === "object" &&
+      "code" in err &&
+      (err as NodeJS.ErrnoException).code === "ENOENT",
+  );
+}
+
+function errorMessage(err: unknown): string {
+  return err instanceof Error ? err.message : String(err);
+}
+
+export type { RuntimeConfig, RuntimeConfigResult };

apps/desktop/src/main/updater.ts

@@ -0,0 +1,100 @@
+import { autoUpdater } from "electron-updater";
+import { app, BrowserWindow, ipcMain } from "electron";
+
+autoUpdater.autoDownload = false;
+autoUpdater.autoInstallOnAppQuit = true;
+
+// Windows arm64 ships its own update metadata channel because
+// electron-builder's `latest.yml` is not arch-suffixed on Windows — both
+// arches would otherwise collide on the same file in the GitHub Release.
+// See scripts/package.mjs (builderArgsForTarget) for the publish-side half
+// of this pact. Pin the channel here so arm64 clients fetch
+// `latest-arm64.yml` instead of the x64 metadata.
+if (process.platform === "win32" && process.arch === "arm64") {
+  autoUpdater.channel = "latest-arm64";
+}
+
+const STARTUP_CHECK_DELAY_MS = 5_000;
+const PERIODIC_CHECK_INTERVAL_MS = 60 * 60 * 1000; // 1 hour
+
+export type ManualUpdateCheckResult =
+  | {
+      ok: true;
+      currentVersion: string;
+      latestVersion: string;
+      available: boolean;
+    }
+  | { ok: false; error: string };
+
+export function setupAutoUpdater(getMainWindow: () => BrowserWindow | null): void {
+  autoUpdater.on("update-available", (info) => {
+    const win = getMainWindow();
+    win?.webContents.send("updater:update-available", {
+      version: info.version,
+      releaseNotes: info.releaseNotes,
+    });
+  });
+
+  autoUpdater.on("download-progress", (progress) => {
+    const win = getMainWindow();
+    win?.webContents.send("updater:download-progress", {
+      percent: progress.percent,
+    });
+  });
+
+  autoUpdater.on("update-downloaded", () => {
+    const win = getMainWindow();
+    win?.webContents.send("updater:update-downloaded");
+  });
+
+  autoUpdater.on("error", (err) => {
+    console.error("Auto-updater error:", err);
+  });
+
+  ipcMain.handle("updater:download", () => {
+    return autoUpdater.downloadUpdate();
+  });
+
+  ipcMain.handle("updater:install", () => {
+    autoUpdater.quitAndInstall(false, true);
+  });
+
+  ipcMain.handle("updater:check", async (): Promise<ManualUpdateCheckResult> => {
+    try {
+      const result = await autoUpdater.checkForUpdates();
+      const currentVersion = app.getVersion();
+      // Trust electron-updater's own decision rather than re-deriving it from
+      // a version-string compare. The two diverge for pre-release channels,
+      // staged rollouts, downgrades, and minimum-system-version gates — in
+      // those cases updateInfo.version differs from app.getVersion() but no
+      // `update-available` event fires, so showing "available" here would
+      // promise a download prompt that never appears.
+      return {
+        ok: true,
+        currentVersion,
+        latestVersion: result?.updateInfo.version ?? currentVersion,
+        available: result?.isUpdateAvailable ?? false,
+      };
+    } catch (err) {
+      return {
+        ok: false,
+        error: err instanceof Error ? err.message : String(err),
+      };
+    }
+  });
+
+  // Initial check shortly after startup so we don't block boot.
+  setTimeout(() => {
+    autoUpdater.checkForUpdates().catch((err) => {
+      console.error("Failed to check for updates:", err);
+    });
+  }, STARTUP_CHECK_DELAY_MS);
+
+  // Background poll so long-running sessions still pick up new releases
+  // without requiring the user to restart the app.
+  setInterval(() => {
+    autoUpdater.checkForUpdates().catch((err) => {
+      console.error("Periodic update check failed:", err);
+    });
+  }, PERIODIC_CHECK_INTERVAL_MS);
+}

apps/desktop/src/main/version-decision.test.ts

@@ -0,0 +1,88 @@
+import { describe, it, expect } from "vitest";
+import { decideVersionAction } from "./version-decision";
+
+describe("decideVersionAction", () => {
+  it("returns not_running when health payload is null", () => {
+    expect(decideVersionAction("v1.0.0", null)).toBe("not_running");
+  });
+
+  it("returns not_running when status is not 'running'", () => {
+    expect(
+      decideVersionAction("v1.0.0", { status: "stopped", cli_version: "v1.0.0" }),
+    ).toBe("not_running");
+  });
+
+  it("returns ok when bundled version is unknown (fail safe)", () => {
+    expect(
+      decideVersionAction(null, {
+        status: "running",
+        cli_version: "v1.0.0",
+        active_task_count: 0,
+      }),
+    ).toBe("ok");
+  });
+
+  it("returns ok when running daemon does not report cli_version (older daemon)", () => {
+    expect(
+      decideVersionAction("v1.0.0", {
+        status: "running",
+        active_task_count: 0,
+      }),
+    ).toBe("ok");
+  });
+
+  it("returns ok when versions match exactly", () => {
+    expect(
+      decideVersionAction("v1.2.3", {
+        status: "running",
+        cli_version: "v1.2.3",
+        active_task_count: 5,
+      }),
+    ).toBe("ok");
+  });
+
+  it("returns restart when versions differ and daemon is idle", () => {
+    expect(
+      decideVersionAction("v1.2.3", {
+        status: "running",
+        cli_version: "v1.2.2",
+        active_task_count: 0,
+      }),
+    ).toBe("restart");
+  });
+
+  it("treats missing active_task_count as 0 (old daemon that still reports cli_version)", () => {
+    expect(
+      decideVersionAction("v1.2.3", {
+        status: "running",
+        cli_version: "v1.2.2",
+      }),
+    ).toBe("restart");
+  });
+
+  it("returns defer when versions differ but daemon is busy", () => {
+    expect(
+      decideVersionAction("v1.2.3", {
+        status: "running",
+        cli_version: "v1.2.2",
+        active_task_count: 2,
+      }),
+    ).toBe("defer");
+  });
+
+  it("transitions defer → restart as tasks drain", () => {
+    // Same bundled version across three observations while the daemon ages.
+    const bundled = "v2.0.0";
+    const base = { status: "running", cli_version: "v1.9.0" } as const;
+
+    expect(
+      decideVersionAction(bundled, { ...base, active_task_count: 3 }),
+    ).toBe("defer");
+    expect(
+      decideVersionAction(bundled, { ...base, active_task_count: 1 }),
+    ).toBe("defer");
+    expect(
+      decideVersionAction(bundled, { ...base, active_task_count: 0 }),
+    ).toBe("restart");
+  });
+});

apps/desktop/src/main/version-decision.ts

@@ -0,0 +1,37 @@
+// Pure decision logic for the daemon version-check flow. Kept in its own
+// module so it can be unit-tested without mocking Electron, execFile, or
+// the HTTP health probe.
+
+export interface VersionCheckHealth {
+  status?: string;
+  cli_version?: string;
+  active_task_count?: number;
+}
+
+export type VersionAction = "restart" | "defer" | "ok" | "not_running";
+
+/**
+ * Decides what the daemon-manager should do given the currently-resolved
+ * bundled CLI version and the latest /health payload.
+ *
+ *   not_running: no daemon is up, nothing to do
+ *   ok:          versions match, OR either side is unknown (fail safe)
+ *   defer:       versions differ but the daemon is busy — wait for drain
+ *   restart:     versions differ and the daemon is idle — safe to restart
+ *
+ * Pure function: no I/O, no side effects, no module state.
+ */
+export function decideVersionAction(
+  bundled: string | null,
+  running: VersionCheckHealth | null,
+): VersionAction {
+  if (!running || running.status !== "running") return "not_running";
+
+  const runningVersion = running.cli_version;
+  if (!bundled || !runningVersion) return "ok";
+  if (runningVersion === bundled) return "ok";
+
+  const activeTasks = running.active_task_count ?? 0;
+  if (activeTasks > 0) return "defer";
+  return "restart";
+}

apps/desktop/src/preload/index.d.ts

@@ -0,0 +1,102 @@
+import { ElectronAPI } from "@electron-toolkit/preload";
+import type { RuntimeConfigResult } from "../shared/runtime-config";
+
+interface DesktopAPI {
+  /** App version + normalized OS, captured synchronously at preload time. */
+  appInfo: {
+    version: string;
+    os: "macos" | "windows" | "linux" | "unknown";
+  };
+  /** OS-preferred locale (BCP 47) injected by main via additionalArguments. */
+  systemLocale: string;
+  /** Subscribe to OS language changes detected after boot. Returns an unsubscribe function. */
+  onSystemLocaleChanged: (callback: (locale: string) => void) => () => void;
+  /** Validated runtime endpoint config, or a blocking config error. */
+  runtimeConfig: RuntimeConfigResult;
+  /** Listen for auth token delivered via deep link. Returns an unsubscribe function. */
+  onAuthToken: (callback: (token: string) => void) => () => void;
+  /** Listen for invitation IDs delivered via deep link. Returns an unsubscribe function. */
+  onInviteOpen: (callback: (invitationId: string) => void) => () => void;
+  /** Open a URL in the default browser. */
+  openExternal: (url: string) => Promise<void>;
+  /** Hide macOS traffic lights for full-screen modals; restore when false. */
+  setImmersiveMode: (immersive: boolean) => Promise<void>;
+  /** Show a native OS notification for a new inbox item. */
+  showNotification: (payload: {
+    slug: string;
+    itemId: string;
+    issueKey: string;
+    title: string;
+    body: string;
+  }) => void;
+  /** Update the OS dock / taskbar unread badge. Pass 0 to clear. */
+  setUnreadBadge: (count: number) => void;
+  /** Listen for "open inbox row" requests from notification clicks. Returns an unsubscribe function. */
+  onInboxOpen: (
+    callback: (payload: {
+      slug: string;
+      itemId: string;
+      issueKey: string;
+    }) => void,
+  ) => () => void;
+}
+
+interface DaemonStatus {
+  state: "running" | "stopped" | "starting" | "stopping" | "installing_cli" | "cli_not_found";
+  pid?: number;
+  uptime?: string;
+  daemonId?: string;
+  deviceName?: string;
+  agents?: string[];
+  workspaceCount?: number;
+  profile?: string;
+  serverUrl?: string;
+}
+
+interface DaemonPrefs {
+  autoStart: boolean;
+  autoStop: boolean;
+}
+
+interface DaemonAPI {
+  start: () => Promise<{ success: boolean; error?: string }>;
+  stop: () => Promise<{ success: boolean; error?: string }>;
+  restart: () => Promise<{ success: boolean; error?: string }>;
+  getStatus: () => Promise<DaemonStatus>;
+  onStatusChange: (callback: (status: DaemonStatus) => void) => () => void;
+  setTargetApiUrl: (url: string) => Promise<void>;
+  syncToken: (token: string, userId: string) => Promise<void>;
+  clearToken: () => Promise<void>;
+  isCliInstalled: () => Promise<boolean>;
+  getPrefs: () => Promise<DaemonPrefs>;
+  setPrefs: (prefs: Partial<DaemonPrefs>) => Promise<DaemonPrefs>;
+  autoStart: () => Promise<void>;
+  retryInstall: () => Promise<void>;
+  startLogStream: () => void;
+  stopLogStream: () => void;
+  onLogLine: (callback: (line: string) => void) => () => void;
+  openLogFile: () => Promise<{ success: boolean; error?: string }>;
+}
+
+interface UpdaterAPI {
+  onUpdateAvailable: (callback: (info: { version: string; releaseNotes?: string }) => void) => () => void;
+  onDownloadProgress: (callback: (progress: { percent: number }) => void) => () => void;
+  onUpdateDownloaded: (callback: () => void) => () => void;
+  downloadUpdate: () => Promise<void>;
+  installUpdate: () => Promise<void>;
+  checkForUpdates: () => Promise<
+    | { ok: true; currentVersion: string; latestVersion: string; available: boolean }
+    | { ok: false; error: string }
+  >;
+}
+
+declare global {
+  interface Window {
+    electron: ElectronAPI;
+    desktopAPI: DesktopAPI;
+    daemonAPI: DaemonAPI;
+    updater: UpdaterAPI;
+  }
+}
+
+export {};

apps/desktop/src/preload/index.ts

@@ -0,0 +1,232 @@
+import { contextBridge, ipcRenderer } from "electron";
+import { electronAPI } from "@electron-toolkit/preload";
+import type { RuntimeConfigResult } from "../shared/runtime-config";
+
+// Synchronously fetch app metadata from main at preload time so the renderer
+// can pass it into CoreProvider during the initial render — the alternative
+// (async ipc.invoke) would race the ApiClient construction in initCore and
+// the first few HTTP requests would go out without X-Client-Version/OS.
+function fetchAppInfo(): { version: string; os: "macos" | "windows" | "linux" | "unknown" } {
+  try {
+    const info = ipcRenderer.sendSync("app:get-info") as
+      | { version: string; os: "macos" | "windows" | "linux" | "unknown" }
+      | undefined;
+    if (info && typeof info.version === "string" && typeof info.os === "string") return info;
+  } catch {
+    // fall through
+  }
+  // Fallback: derive OS from process.platform; version unknown.
+  const p = process.platform;
+  const os: "macos" | "windows" | "linux" | "unknown" =
+    p === "darwin" ? "macos" : p === "win32" ? "windows" : p === "linux" ? "linux" : "unknown";
+  return { version: "unknown", os };
+}
+
+function fetchRuntimeConfig(): RuntimeConfigResult {
+  try {
+    const result = ipcRenderer.sendSync("runtime-config:get") as RuntimeConfigResult | undefined;
+    if (result && typeof result === "object" && "ok" in result) return result;
+  } catch (err) {
+    return {
+      ok: false,
+      error: {
+        message: err instanceof Error ? err.message : String(err),
+      },
+    };
+  }
+  return { ok: false, error: { message: "Runtime config unavailable" } };
+}
+
+const appInfo = fetchAppInfo();
+const runtimeConfig = fetchRuntimeConfig();
+
+// Read the OS-preferred locale that main injected via additionalArguments.
+// Zero IPC, zero blocking — process.argv is populated before preload runs.
+function fetchSystemLocale(): string {
+  const arg = process.argv.find((a) => a.startsWith("--multica-locale="));
+  return arg?.split("=")[1] ?? "en";
+}
+
+const systemLocale = fetchSystemLocale();
+
+const desktopAPI = {
+  /** App version + normalized OS. Read once at preload time so the renderer
+   *  can use it synchronously when initializing the API client. */
+  appInfo,
+  /** OS-preferred locale (BCP 47), passed from main via additionalArguments.
+   *  Used by the renderer's LocaleAdapter as the system-preference signal. */
+  systemLocale,
+  /** Subscribe to OS language changes detected after boot. The renderer
+   *  decides whether to act (no-op when the user has an explicit Settings
+   *  choice). Returns an unsubscribe function. */
+  onSystemLocaleChanged: (callback: (locale: string) => void) => {
+    const handler = (_event: Electron.IpcRendererEvent, locale: string) =>
+      callback(locale);
+    ipcRenderer.on("locale:system-changed", handler);
+    return () => {
+      ipcRenderer.removeListener("locale:system-changed", handler);
+    };
+  },
+  /** Validated runtime endpoint config, or a blocking config error. */
+  runtimeConfig,
+  /** Listen for auth token delivered via deep link */
+  onAuthToken: (callback: (token: string) => void) => {
+    const handler = (_event: Electron.IpcRendererEvent, token: string) =>
+      callback(token);
+    ipcRenderer.on("auth:token", handler);
+    return () => {
+      ipcRenderer.removeListener("auth:token", handler);
+    };
+  },
+  /** Listen for invitation IDs delivered via deep link */
+  onInviteOpen: (callback: (invitationId: string) => void) => {
+    const handler = (_event: Electron.IpcRendererEvent, invitationId: string) =>
+      callback(invitationId);
+    ipcRenderer.on("invite:open", handler);
+    return () => {
+      ipcRenderer.removeListener("invite:open", handler);
+    };
+  },
+  /** Open a URL in the default browser */
+  openExternal: (url: string) => ipcRenderer.invoke("shell:openExternal", url),
+  /** Toggle immersive mode — hide macOS traffic lights for full-screen modals */
+  setImmersiveMode: (immersive: boolean) =>
+    ipcRenderer.invoke("window:setImmersive", immersive),
+  /**
+   * Show a native OS notification for a new inbox item. Fired from the
+   * renderer only when the app is unfocused — in-focus feedback is the
+   * inbox sidebar's unread styling. `slug`, `itemId`, and `issueKey` are
+   * all round-tripped on click: slug pins routing to the source workspace
+   * (the user may switch workspaces before clicking the banner), itemId
+   * lets the renderer mark the row read, issueKey maps to the inbox URL
+   * param.
+   */
+  showNotification: (payload: {
+    slug: string;
+    itemId: string;
+    issueKey: string;
+    title: string;
+    body: string;
+  }) => ipcRenderer.send("notification:show", payload),
+  /**
+   * Update the OS dock / taskbar unread badge. Pass 0 to clear. Values
+   * above 99 render as "99+" (capping is handled in the main process).
+   */
+  setUnreadBadge: (count: number) =>
+    ipcRenderer.send("badge:set", Math.max(0, Math.floor(count))),
+  /**
+   * Subscribe to "open this inbox row" requests sent by the main process
+   * when the user clicks an OS notification banner. Returns an unsubscribe
+   * function. The payload echoes the `slug`, `itemId`, and `issueKey` that
+   * were passed to `showNotification`.
+   */
+  onInboxOpen: (
+    callback: (payload: {
+      slug: string;
+      itemId: string;
+      issueKey: string;
+    }) => void,
+  ) => {
+    const handler = (
+      _event: Electron.IpcRendererEvent,
+      payload: { slug: string; itemId: string; issueKey: string },
+    ) => callback(payload);
+    ipcRenderer.on("inbox:open", handler);
+    return () => {
+      ipcRenderer.removeListener("inbox:open", handler);
+    };
+  },
+};
+
+interface DaemonStatus {
+  state: "running" | "stopped" | "starting" | "stopping" | "installing_cli" | "cli_not_found";
+  pid?: number;
+  uptime?: string;
+  daemonId?: string;
+  deviceName?: string;
+  agents?: string[];
+  workspaceCount?: number;
+  profile?: string;
+  serverUrl?: string;
+}
+
+const daemonAPI = {
+  start: (): Promise<{ success: boolean; error?: string }> =>
+    ipcRenderer.invoke("daemon:start"),
+  stop: (): Promise<{ success: boolean; error?: string }> =>
+    ipcRenderer.invoke("daemon:stop"),
+  restart: (): Promise<{ success: boolean; error?: string }> =>
+    ipcRenderer.invoke("daemon:restart"),
+  getStatus: (): Promise<DaemonStatus> =>
+    ipcRenderer.invoke("daemon:get-status"),
+  onStatusChange: (callback: (status: DaemonStatus) => void) => {
+    const handler = (_: unknown, status: DaemonStatus) => callback(status);
+    ipcRenderer.on("daemon:status", handler);
+    return () => ipcRenderer.removeListener("daemon:status", handler);
+  },
+  setTargetApiUrl: (url: string): Promise<void> =>
+    ipcRenderer.invoke("daemon:set-target-api-url", url),
+  syncToken: (token: string, userId: string): Promise<void> =>
+    ipcRenderer.invoke("daemon:sync-token", token, userId),
+  clearToken: (): Promise<void> =>
+    ipcRenderer.invoke("daemon:clear-token"),
+  isCliInstalled: (): Promise<boolean> =>
+    ipcRenderer.invoke("daemon:is-cli-installed"),
+  getPrefs: (): Promise<{ autoStart: boolean; autoStop: boolean }> =>
+    ipcRenderer.invoke("daemon:get-prefs"),
+  setPrefs: (prefs: Partial<{ autoStart: boolean; autoStop: boolean }>): Promise<{ autoStart: boolean; autoStop: boolean }> =>
+    ipcRenderer.invoke("daemon:set-prefs", prefs),
+  autoStart: (): Promise<void> =>
+    ipcRenderer.invoke("daemon:auto-start"),
+  retryInstall: (): Promise<void> =>
+    ipcRenderer.invoke("daemon:retry-install"),
+  startLogStream: () => ipcRenderer.send("daemon:start-log-stream"),
+  stopLogStream: () => ipcRenderer.send("daemon:stop-log-stream"),
+  onLogLine: (callback: (line: string) => void) => {
+    const handler = (_: unknown, line: string) => callback(line);
+    ipcRenderer.on("daemon:log-line", handler);
+    return () => ipcRenderer.removeListener("daemon:log-line", handler);
+  },
+  openLogFile: (): Promise<{ success: boolean; error?: string }> =>
+    ipcRenderer.invoke("daemon:open-log-file"),
+};
+
+const updaterAPI = {
+  onUpdateAvailable: (callback: (info: { version: string; releaseNotes?: string }) => void) => {
+    const handler = (_: unknown, info: { version: string; releaseNotes?: string }) => callback(info);
+    ipcRenderer.on("updater:update-available", handler);
+    return () => ipcRenderer.removeListener("updater:update-available", handler);
+  },
+  onDownloadProgress: (callback: (progress: { percent: number }) => void) => {
+    const handler = (_: unknown, progress: { percent: number }) => callback(progress);
+    ipcRenderer.on("updater:download-progress", handler);
+    return () => ipcRenderer.removeListener("updater:download-progress", handler);
+  },
+  onUpdateDownloaded: (callback: () => void) => {
+    const handler = () => callback();
+    ipcRenderer.on("updater:update-downloaded", handler);
+    return () => ipcRenderer.removeListener("updater:update-downloaded", handler);
+  },
+  downloadUpdate: () => ipcRenderer.invoke("updater:download"),
+  installUpdate: () => ipcRenderer.invoke("updater:install"),
+  checkForUpdates: (): Promise<
+    | { ok: true; currentVersion: string; latestVersion: string; available: boolean }
+    | { ok: false; error: string }
+  > => ipcRenderer.invoke("updater:check"),
+};
+
+if (process.contextIsolated) {
+  contextBridge.exposeInMainWorld("electron", electronAPI);
+  contextBridge.exposeInMainWorld("desktopAPI", desktopAPI);
+  contextBridge.exposeInMainWorld("daemonAPI", daemonAPI);
+  contextBridge.exposeInMainWorld("updater", updaterAPI);
+} else {
+  // @ts-expect-error - fallback for non-isolated context
+  window.electron = electronAPI;
+  // @ts-expect-error - fallback for non-isolated context
+  window.desktopAPI = desktopAPI;
+  // @ts-expect-error - fallback for non-isolated context
+  window.daemonAPI = daemonAPI;
+  // @ts-expect-error - fallback for non-isolated context
+  window.updater = updaterAPI;
+}

apps/desktop/src/renderer/index.html

@@ -0,0 +1,12 @@
+<!doctype html>
+<html lang="en" class="h-full">
+  <head>
+    <meta charset="UTF-8" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" />
+    <title>Multica</title>
+  </head>
+  <body class="h-full overflow-hidden antialiased font-sans">
+    <div id="root" class="h-full"></div>
+    <script type="module" src="/src/main.tsx"></script>
+  </body>
+</html>

apps/desktop/src/renderer/src/App.tsx

@@ -0,0 +1,334 @@
+import { useEffect, useLayoutEffect, useMemo, useRef, useState } from "react";
+import { useQuery, useQueryClient } from "@tanstack/react-query";
+import { CoreProvider } from "@multica/core/platform";
+import { pickLocale } from "@multica/core/i18n";
+import { useAuthStore } from "@multica/core/auth";
+import { workspaceKeys, workspaceListOptions } from "@multica/core/workspace/queries";
+import { api } from "@multica/core/api";
+import { useHasOnboarded } from "@multica/core/paths";
+import { ThemeProvider } from "@multica/ui/components/common/theme-provider";
+import { MulticaIcon } from "@multica/ui/components/common/multica-icon";
+import { Toaster } from "@multica/ui/components/ui/sonner";
+import { DesktopLoginPage } from "./pages/login";
+import { DesktopShell } from "./components/desktop-layout";
+import { PageviewTracker } from "./components/pageview-tracker";
+import { UpdateNotification } from "./components/update-notification";
+import { useTabStore } from "./stores/tab-store";
+import { useWindowOverlayStore } from "./stores/window-overlay-store";
+import { useDaemonIPCBridge } from "./platform/daemon-ipc-bridge";
+import { createDesktopLocaleAdapter } from "./platform/i18n-adapter";
+import { RESOURCES } from "@multica/views/locales";
+
+
+function AppContent() {
+  const user = useAuthStore((s) => s.user);
+  const isLoading = useAuthStore((s) => s.isLoading);
+  const qc = useQueryClient();
+  // Deep-link login runs loginWithToken → syncToken → listWorkspaces →
+  // setQueryData sequentially. loginWithToken sets user+isLoading=false
+  // as soon as getMe resolves, which would cause DesktopShell to mount
+  // before the workspace list is hydrated and briefly see `!workspace`.
+  // This local flag keeps the loading screen up until the whole chain
+  // finishes, so IndexRedirect gets a definitive workspace state on
+  // first render.
+  const [bootstrapping, setBootstrapping] = useState(false);
+
+  const runtimeConfig = window.desktopAPI.runtimeConfig.ok
+    ? window.desktopAPI.runtimeConfig.config
+    : null;
+
+  // Tell the main process which backend URL we talk to, so daemon-manager
+  // can pick the matching CLI profile (server_url from ~/.multica config).
+  useEffect(() => {
+    if (!runtimeConfig) return;
+    window.daemonAPI.setTargetApiUrl(runtimeConfig.apiUrl);
+  }, [runtimeConfig]);
+
+  // Listen for invite IDs delivered via deep link (multica://invite/<id>).
+  // We open the overlay regardless of login state — if the user isn't logged
+  // in, InvitePage's queries will fail and render the "not found" state,
+  // which is acceptable; the expected pre-flight happens in the web app
+  // (login + next=/invite/... dance) before the deep link is ever dispatched.
+  useEffect(() => {
+    return window.desktopAPI.onInviteOpen((invitationId) => {
+      useWindowOverlayStore.getState().open({ type: "invite", invitationId });
+    });
+  }, []);
+
+  // Listen for auth token delivered via deep link (multica://auth/callback?token=...).
+  // daemonAPI.syncToken is handled separately by the [user] effect below, which
+  // fires whenever a user logs in (deep link, session restore, account switch).
+  useEffect(() => {
+    return window.desktopAPI.onAuthToken(async (token) => {
+      setBootstrapping(true);
+      try {
+        await useAuthStore.getState().loginWithToken(token);
+        // Seed React Query cache with the workspace list so the index-route
+        // redirect (routes.tsx `IndexRedirect`) can resolve the initial
+        // destination without a second fetch. Workspace side-effects
+        // (setCurrentWorkspace, persist namespace) are synced later by
+        // WorkspaceRouteLayout when the URL resolves.
+        const wsList = await api.listWorkspaces();
+        qc.setQueryData(workspaceKeys.list(), wsList);
+      } catch {
+        // Token invalid or expired — user stays on login page
+      } finally {
+        setBootstrapping(false);
+      }
+    });
+  }, [qc]);
+
+  // Sync token and start the daemon whenever the user logs in.
+  useEffect(() => {
+    if (!user) return;
+    const token = localStorage.getItem("multica_token");
+    if (!token) return;
+    const userId = user.id;
+    (async () => {
+      try {
+        await window.daemonAPI.syncToken(token, userId);
+        await window.daemonAPI.autoStart();
+      } catch (err) {
+        console.error("Failed to sync daemon on login", err);
+      }
+    })();
+  }, [user]);
+
+  // When a user who started the session with zero workspaces creates their
+  // first one, restart the daemon so it picks up the new workspace
+  // immediately (otherwise workspaceSyncLoop's next 30s tick would be the
+  // earliest pickup point). Specifically scoped to "started empty" because
+  // account switches (user A logout → user B login) should not trigger a
+  // daemon restart here — daemon-manager already restarts on user change
+  // via syncToken.
+  const { data: workspaces = [], isFetched: workspaceListFetched } = useQuery({
+    ...workspaceListOptions(),
+    enabled: !!user,
+  });
+  const wsCount = workspaces.length;
+  const hasOnboarded = useHasOnboarded();
+
+  // Bridge local daemon IPC status into the runtimes cache so this user's
+  // own daemon flips to offline/online sub-second instead of waiting on the
+  // server's 75s sweeper. Resolves wsId from the active tab so workspace
+  // switches automatically rebind the subscription.
+  const activeWorkspaceSlug = useTabStore((s) => s.activeWorkspaceSlug);
+  const activeWsId = activeWorkspaceSlug
+    ? workspaces.find((w) => w.slug === activeWorkspaceSlug)?.id
+    : undefined;
+  useDaemonIPCBridge(activeWsId);
+
+  // Pre-workspace overlay routing for desktop. Mirrors the web entry-point
+  // judgment in callback / login:
+  //   un-onboarded:
+  //     pending invites on email → /invitations overlay
+  //     no invites               → /onboarding overlay
+  //   already onboarded:
+  //     zero workspaces          → /workspaces/new overlay
+  //     ≥1 workspaces            → no overlay, fall through to dashboard
+  //
+  // The "un-onboarded but in workspace" state is now physically impossible
+  // because backend transactions atomically set onboarded_at when a user
+  // joins the `member` table. Anyone with workspaces is by definition
+  // onboarded.
+  useEffect(() => {
+    if (!user || !workspaceListFetched) return undefined;
+    const { overlay, open } = useWindowOverlayStore.getState();
+    if (overlay) return undefined;
+    if (wsCount > 0) return undefined;
+    if (!hasOnboarded) {
+      // Look up pending invitations by email. Network blip is non-fatal —
+      // fall through to onboarding so the user isn't stuck on a blank
+      // window. The sidebar's pending-invitations dropdown will surface
+      // missed invites later once they're onboarded.
+      let cancelled = false;
+      void api
+        .listMyInvitations()
+        .then((invites) => {
+          if (cancelled) return;
+          const { overlay: latestOverlay, open: latestOpen } =
+            useWindowOverlayStore.getState();
+          if (latestOverlay) return;
+          if (invites.length > 0) {
+            qc.setQueryData(workspaceKeys.myInvitations(), invites);
+            latestOpen({ type: "invitations" });
+          } else {
+            latestOpen({ type: "onboarding" });
+          }
+        })
+        .catch(() => {
+          if (cancelled) return;
+          const { overlay: latestOverlay, open: latestOpen } =
+            useWindowOverlayStore.getState();
+          if (latestOverlay) return;
+          latestOpen({ type: "onboarding" });
+        });
+      return () => {
+        cancelled = true;
+      };
+    }
+    open({ type: "new-workspace" });
+    return undefined;
+  }, [user, workspaceListFetched, wsCount, workspaces, hasOnboarded, qc]);
+
+  // Validate persisted tab state against the current user's workspace list,
+  // and pick an active workspace if none is set. Runs in useLayoutEffect
+  // (synchronously after render, before paint) rather than the render
+  // phase — the original render-phase pattern triggered React's
+  // "Cannot update a component while rendering a different component"
+  // warning because `switchWorkspace` is a Zustand setState that the
+  // TabBar is subscribed to. useLayoutEffect flushes both renders before
+  // the user sees anything, so there's no visible flicker.
+  //
+  // Gate on `workspaceListFetched`: useQuery defaults `data` to `[]` before
+  // the first fetch, so without this guard we'd run validation against an
+  // empty slug set, wipe the persisted `activeWorkspaceSlug`, then fall
+  // back to `workspaces[0]` once the real list arrives — losing the user's
+  // last-opened workspace on every app start.
+  useLayoutEffect(() => {
+    if (!workspaceListFetched) return;
+    const validSlugs = new Set(workspaces.map((w) => w.slug));
+    useTabStore.getState().validateWorkspaceSlugs(validSlugs);
+    const { activeWorkspaceSlug, switchWorkspace } = useTabStore.getState();
+    if (!activeWorkspaceSlug && workspaces.length > 0) {
+      switchWorkspace(workspaces[0].slug);
+    }
+  }, [workspaces, workspaceListFetched]);
+
+  // null = undecided (pre-login or list hasn't settled yet)
+  // true  = session started with zero workspaces; next transition to >=1 triggers restart
+  // false = session started with >=1 workspace, OR we've already restarted; skip
+  const sessionStartedEmptyRef = useRef<boolean | null>(null);
+  useEffect(() => {
+    if (!user) {
+      sessionStartedEmptyRef.current = null;
+      return;
+    }
+    if (!workspaceListFetched) return;
+    if (sessionStartedEmptyRef.current === null) {
+      sessionStartedEmptyRef.current = wsCount === 0;
+      return;
+    }
+    if (sessionStartedEmptyRef.current && wsCount >= 1) {
+      void window.daemonAPI.restart();
+      sessionStartedEmptyRef.current = false;
+    }
+  }, [user, workspaceListFetched, wsCount]);
+
+  if (isLoading || bootstrapping) {
+    return (
+      <div className="flex h-screen items-center justify-center">
+        <MulticaIcon className="size-6 animate-pulse" />
+      </div>
+    );
+  }
+
+  // Pageview tracker sits at the app root so it covers every visible
+  // surface (login, overlays, tab paths) — mounting it inside DesktopShell
+  // would miss the logged-out and overlay states.
+  return (
+    <>
+      <PageviewTracker />
+      {user ? <DesktopShell /> : <DesktopLoginPage />}
+    </>
+  );
+}
+
+function BlockingRuntimeConfigError({ message }: { message: string }) {
+  return (
+    <div className="flex h-screen items-center justify-center bg-background p-8 text-foreground">
+      <div className="max-w-xl rounded-lg border bg-card p-6 shadow-sm">
+        <h1 className="text-lg font-semibold">Desktop configuration error</h1>
+        <p className="mt-3 text-sm text-muted-foreground">
+          Multica Desktop could not load <code>~/.multica/desktop.json</code>. Fix or remove the file and restart the app.
+        </p>
+        <pre className="mt-4 whitespace-pre-wrap rounded-md bg-muted p-3 text-xs text-muted-foreground">
+          {message}
+        </pre>
+      </div>
+    </div>
+  );
+}
+
+// On logout, wipe desktop-only in-memory state and stop the daemon so that
+// a subsequent login as a different user never inherits the previous user's
+// tabs, overlay, or credentials. Zustand persist only writes to localStorage;
+// useLogout clears the storage key, but the live stores stay populated until
+// we explicitly reset them here.
+async function handleDaemonLogout() {
+  useTabStore.getState().reset();
+  useWindowOverlayStore.getState().close();
+  try {
+    await window.daemonAPI.clearToken();
+  } catch {
+    // Best-effort — clearing is followed by stop which also hardens state.
+  }
+  try {
+    await window.daemonAPI.stop();
+  } catch {
+    // Daemon may already be stopped.
+  }
+}
+
+export default function App() {
+  const { version, os } = window.desktopAPI.appInfo;
+  const systemLocale = window.desktopAPI.systemLocale;
+  const runtimeConfigResult = window.desktopAPI.runtimeConfig;
+  // Stable identity reference so downstream effects (WS reconnect) don't
+  // tear down on every parent render.
+  const identity = useMemo(
+    () => ({ platform: "desktop", version, os }),
+    [version, os],
+  );
+  // Locale resolution happens once at app boot. Switching language goes
+  // through window.location.reload() to avoid hydration mismatch.
+  const localeAdapter = useMemo(
+    () => createDesktopLocaleAdapter(systemLocale),
+    [systemLocale],
+  );
+  const locale = useMemo(() => pickLocale(localeAdapter), [localeAdapter]);
+  const resources = useMemo(
+    () => ({ [locale]: RESOURCES[locale] }),
+    [locale],
+  );
+
+  // React to OS-level language changes detected by main on focus regain.
+  // Only act when the user is following the system signal (no explicit
+  // Settings choice) — otherwise their preference wins. Cross-device sync
+  // for the explicit-choice case is handled inside CoreProvider.
+  useEffect(() => {
+    return window.desktopAPI.onSystemLocaleChanged((nextSystemLocale) => {
+      if (localeAdapter.getUserChoice()) return;
+      const next = pickLocale({
+        ...localeAdapter,
+        getSystemPreferences: () =>
+          nextSystemLocale ? [nextSystemLocale] : [],
+      });
+      if (next === locale) return;
+      localeAdapter.persist(next);
+      window.location.reload();
+    });
+  }, [localeAdapter, locale]);
+
+  return (
+    <ThemeProvider>
+      {runtimeConfigResult.ok ? (
+        <CoreProvider
+          apiBaseUrl={runtimeConfigResult.config.apiUrl}
+          wsUrl={runtimeConfigResult.config.wsUrl}
+          onLogout={handleDaemonLogout}
+          identity={identity}
+          locale={locale}
+          resources={resources}
+          localeAdapter={localeAdapter}
+        >
+          <AppContent />
+        </CoreProvider>
+      ) : (
+        <BlockingRuntimeConfigError message={runtimeConfigResult.error.message} />
+      )}
+      <Toaster />
+      <UpdateNotification />
+    </ThemeProvider>
+  );
+}

apps/desktop/src/renderer/src/components/daemon-panel.tsx

@@ -0,0 +1,675 @@
+import {
+  Fragment,
+  useCallback,
+  useEffect,
+  useMemo,
+  useRef,
+  useState,
+  type ReactNode,
+} from "react";
+import {
+  ArrowDown,
+  Copy as CopyIcon,
+  Search,
+  Server,
+  Trash2,
+  X,
+} from "lucide-react";
+import { cn } from "@multica/ui/lib/utils";
+import { Button } from "@multica/ui/components/ui/button";
+import {
+  Dialog,
+  DialogContent,
+  DialogTitle,
+} from "@multica/ui/components/ui/dialog";
+import { toast } from "sonner";
+import type { DaemonStatus } from "../../../shared/daemon-types";
+import {
+  DAEMON_STATE_COLORS,
+  DAEMON_STATE_LABELS,
+  formatUptime,
+} from "../../../shared/daemon-types";
+import { parseLogLine, type LogLevel, type ParsedLogLine } from "./parse-daemon-log";
+
+interface DaemonPanelProps {
+  open: boolean;
+  onOpenChange: (open: boolean) => void;
+  status: DaemonStatus;
+  /** Number of runtimes this local daemon has registered (for the context badge). */
+  runtimeCount: number;
+}
+
+const MAX_LOG_LINES = 500;
+const LEVELS: readonly LogLevel[] = ["DEBUG", "INFO", "WARN", "ERROR"];
+
+const LEVEL_BADGE_CLASS: Record<LogLevel, string> = {
+  DEBUG: "border-muted-foreground/25 text-muted-foreground/70",
+  INFO: "border-foreground/15 text-foreground/80",
+  WARN: "border-warning/40 text-warning",
+  ERROR: "border-destructive/40 text-destructive",
+};
+
+// What gets rendered in the viewport — a single line or a folded group of
+// consecutive lines that share the same `message`. The group form is what
+// turns a wall of `DBG poll: no tasks` into a single placeholder.
+type DisplayItem =
+  | { kind: "line"; line: ParsedLogLine }
+  | { kind: "group"; first: ParsedLogLine; rest: ParsedLogLine[] };
+
+export function DaemonPanel({
+  open,
+  onOpenChange,
+  status,
+  runtimeCount,
+}: DaemonPanelProps) {
+  const [logs, setLogs] = useState<ParsedLogLine[]>([]);
+  const [search, setSearch] = useState("");
+  // Each level chip is an independent toggle. DEBUG is off by default so
+  // poll-loop noise doesn't drown out real events when the panel opens —
+  // users opt in if they want to see it.
+  const [enabledLevels, setEnabledLevels] = useState<Set<LogLevel>>(
+    () => new Set<LogLevel>(["INFO", "WARN", "ERROR"]),
+  );
+  const [autoScroll, setAutoScroll] = useState(true);
+  const [expandedFields, setExpandedFields] = useState<Set<number>>(new Set());
+  const [expandedGroups, setExpandedGroups] = useState<Set<number>>(new Set());
+
+  const idCounterRef = useRef(0);
+  const logContainerRef = useRef<HTMLDivElement>(null);
+
+  // --- Log stream subscription ---
+  // Active only while the modal is open. On open we replay the file's tail
+  // (~200 lines) so users have context for "what just happened"; on close
+  // we tear down the watcher so the main process isn't doing work for a
+  // hidden UI.
+  useEffect(() => {
+    if (!open) return;
+    setLogs([]);
+    setExpandedFields(new Set());
+    setExpandedGroups(new Set());
+    idCounterRef.current = 0;
+
+    window.daemonAPI.startLogStream();
+    const unsub = window.daemonAPI.onLogLine((line) => {
+      setLogs((prev) => {
+        const id = ++idCounterRef.current;
+        const parsed = parseLogLine(line, id);
+        const next =
+          prev.length >= MAX_LOG_LINES
+            ? [...prev.slice(prev.length - MAX_LOG_LINES + 1), parsed]
+            : [...prev, parsed];
+        return next;
+      });
+    });
+    return () => {
+      unsub();
+      window.daemonAPI.stopLogStream();
+    };
+  }, [open]);
+
+  // --- Derived: counts per level (for filter chip badges) ---
+  const levelCounts = useMemo(() => {
+    const counts: Record<LogLevel, number> = {
+      DEBUG: 0,
+      INFO: 0,
+      WARN: 0,
+      ERROR: 0,
+    };
+    for (const l of logs) {
+      if (l.level) counts[l.level] += 1;
+    }
+    return counts;
+  }, [logs]);
+
+  // --- Derived: filtered list (level toggle + search) ---
+  // Lines that didn't parse (level = null) always pass — they're typically
+  // panic stack traces / partial writes; never silently drop them.
+  const filtered = useMemo(() => {
+    let result = logs;
+    result = result.filter((l) => {
+      if (!l.level) return true;
+      return enabledLevels.has(l.level);
+    });
+    if (search) {
+      const q = search.toLowerCase();
+      result = result.filter((l) => l.raw.toLowerCase().includes(q));
+    }
+    return result;
+  }, [logs, enabledLevels, search]);
+
+  // --- Derived: collapse runs of consecutive lines that share the same
+  // message into a single group placeholder. The most common case is the
+  // 1-min `DBG poll: no tasks` heartbeat that otherwise pushes real events
+  // off-screen. Grouping happens AFTER filtering so toggling DEBUG off
+  // doesn't strand groups.
+  const displayed = useMemo<DisplayItem[]>(() => {
+    const out: DisplayItem[] = [];
+    for (const line of filtered) {
+      const last = out[out.length - 1];
+      if (!last) {
+        out.push({ kind: "line", line });
+        continue;
+      }
+      const lastMessage =
+        last.kind === "line" ? last.line.message : last.first.message;
+      if (lastMessage && lastMessage === line.message) {
+        if (last.kind === "line") {
+          out[out.length - 1] = {
+            kind: "group",
+            first: last.line,
+            rest: [line],
+          };
+        } else {
+          last.rest.push(line);
+        }
+      } else {
+        out.push({ kind: "line", line });
+      }
+    }
+    return out;
+  }, [filtered]);
+
+  // --- Auto-scroll: pin to bottom while live; release on user scroll ---
+  useEffect(() => {
+    if (!autoScroll) return;
+    const el = logContainerRef.current;
+    if (el) el.scrollTop = el.scrollHeight;
+  }, [displayed, autoScroll]);
+
+  const handleScroll = useCallback(() => {
+    const el = logContainerRef.current;
+    if (!el) return;
+    const atBottom = el.scrollHeight - el.scrollTop - el.clientHeight < 40;
+    // Only flip auto-scroll OFF on user-initiated scroll-up; never flip ON
+    // here. Re-enabling lives in the "Jump to latest" footer button so a
+    // burst of lines doesn't yank a reading user back to the bottom.
+    if (!atBottom && autoScroll) setAutoScroll(false);
+  }, [autoScroll]);
+
+  const handleResume = useCallback(() => {
+    setAutoScroll(true);
+    const el = logContainerRef.current;
+    if (el) el.scrollTop = el.scrollHeight;
+  }, []);
+
+  const handleCopy = useCallback(async () => {
+    const text = filtered.map((l) => l.raw).join("\n");
+    try {
+      await navigator.clipboard.writeText(text);
+      toast.success(
+        `Copied ${filtered.length} line${filtered.length === 1 ? "" : "s"}`,
+      );
+    } catch (err) {
+      toast.error("Failed to copy", {
+        description: err instanceof Error ? err.message : String(err),
+      });
+    }
+  }, [filtered]);
+
+  const handleClear = useCallback(() => {
+    setLogs([]);
+    setExpandedFields(new Set());
+    setExpandedGroups(new Set());
+  }, []);
+
+  const toggleLevel = useCallback((lv: LogLevel) => {
+    setEnabledLevels((prev) => {
+      const next = new Set(prev);
+      if (next.has(lv)) next.delete(lv);
+      else next.add(lv);
+      return next;
+    });
+  }, []);
+
+  const toggleFields = useCallback((id: number) => {
+    setExpandedFields((prev) => {
+      const next = new Set(prev);
+      if (next.has(id)) next.delete(id);
+      else next.add(id);
+      return next;
+    });
+  }, []);
+
+  const toggleGroup = useCallback((id: number) => {
+    setExpandedGroups((prev) => {
+      const next = new Set(prev);
+      if (next.has(id)) next.delete(id);
+      else next.add(id);
+      return next;
+    });
+  }, []);
+
+  const hasActiveFilter = !!search || enabledLevels.size < LEVELS.length;
+
+  return (
+    <Dialog open={open} onOpenChange={onOpenChange}>
+      <DialogContent
+        className="flex h-[85vh] flex-col gap-0 overflow-hidden p-0 sm:max-w-5xl"
+        showCloseButton={false}
+      >
+        {/* Header */}
+        <div className="flex shrink-0 items-center justify-between gap-3 border-b px-4 py-3">
+          <div className="flex min-w-0 items-center gap-2">
+            <Server className="size-4 shrink-0 text-muted-foreground" />
+            <DialogTitle className="text-sm font-medium">
+              Local daemon logs
+            </DialogTitle>
+            <ContextBadge status={status} runtimeCount={runtimeCount} />
+          </div>
+          <button
+            type="button"
+            onClick={() => onOpenChange(false)}
+            aria-label="Close"
+            className="flex size-7 shrink-0 items-center justify-center rounded-md text-muted-foreground transition-colors hover:bg-muted hover:text-foreground focus-visible:outline-none focus-visible:ring-2 focus-visible:ring-ring"
+          >
+            <X className="size-4" />
+          </button>
+        </div>
+
+        {/* Toolbar */}
+        <div className="flex shrink-0 flex-wrap items-center gap-2 border-b px-4 py-2">
+          {/* Search */}
+          <div className="relative w-56">
+            <Search className="pointer-events-none absolute left-2 top-1/2 size-3.5 -translate-y-1/2 text-muted-foreground" />
+            <input
+              value={search}
+              onChange={(e) => setSearch(e.target.value)}
+              placeholder="Search…"
+              className="h-7 w-full rounded-md border bg-background pl-7 pr-2 text-xs placeholder:text-muted-foreground focus:outline-none focus:ring-1 focus:ring-ring"
+            />
+          </div>
+
+          {/* Level toggle chips. Each chip is independent — click to
+              show/hide that level. DEBUG starts hidden because the
+              poll-loop heartbeat dominates otherwise. */}
+          <div className="flex items-center gap-1">
+            {LEVELS.map((lv) => (
+              <FilterChip
+                key={lv}
+                active={enabledLevels.has(lv)}
+                onClick={() => toggleLevel(lv)}
+                label={lv}
+                count={levelCounts[lv]}
+                variant={lv}
+              />
+            ))}
+          </div>
+
+          {/* Right-aligned actions */}
+          <div className="ml-auto flex items-center gap-1">
+            <Button
+              variant="ghost"
+              size="sm"
+              className="h-7"
+              onClick={handleCopy}
+              disabled={filtered.length === 0}
+            >
+              <CopyIcon className="size-3.5 mr-1.5" />
+              Copy
+            </Button>
+            <Button
+              variant="ghost"
+              size="sm"
+              className="h-7"
+              onClick={handleClear}
+              disabled={logs.length === 0}
+            >
+              <Trash2 className="size-3.5 mr-1.5" />
+              Clear
+            </Button>
+          </div>
+        </div>
+
+        {/* Logs viewport */}
+        <div
+          ref={logContainerRef}
+          onScroll={handleScroll}
+          className="min-h-0 flex-1 overflow-y-auto bg-muted/20 px-2 py-1 font-mono text-xs"
+        >
+          {displayed.length === 0 ? (
+            <EmptyState
+              hasLogs={logs.length > 0}
+              hasFilter={hasActiveFilter}
+              isRunning={status.state === "running"}
+            />
+          ) : (
+            <div className="flex flex-col">
+              {displayed.map((item) =>
+                item.kind === "line" ? (
+                  <LogLineRow
+                    key={item.line.id}
+                    line={item.line}
+                    expanded={expandedFields.has(item.line.id)}
+                    onToggle={() => toggleFields(item.line.id)}
+                    search={search}
+                  />
+                ) : (
+                  <GroupRows
+                    key={item.first.id}
+                    first={item.first}
+                    rest={item.rest}
+                    expanded={expandedGroups.has(item.first.id)}
+                    onToggle={() => toggleGroup(item.first.id)}
+                    expandedFields={expandedFields}
+                    onToggleFields={toggleFields}
+                    search={search}
+                  />
+                ),
+              )}
+            </div>
+          )}
+        </div>
+
+        {/* Status bar — count only. The "is the user following" state is
+            communicated implicitly by the presence of the Jump-to-latest
+            button below; an explicit "Paused" word read as "log stream is
+            paused" (it isn't — data keeps flowing into the buffer). */}
+        <div className="flex shrink-0 items-center justify-between border-t bg-muted/30 px-4 py-1.5 text-xs text-muted-foreground">
+          <span className="tabular-nums">
+            Showing {filtered.length} of {logs.length}
+            {logs.length === MAX_LOG_LINES && (
+              <span className="ml-1 text-muted-foreground/60">
+                (buffer full)
+              </span>
+            )}
+          </span>
+          {!autoScroll && (
+            <button
+              type="button"
+              onClick={handleResume}
+              className="inline-flex items-center gap-1 rounded-md px-2 py-0.5 hover:bg-muted hover:text-foreground"
+            >
+              <ArrowDown className="size-3" />
+              Jump to latest
+            </button>
+          )}
+        </div>
+      </DialogContent>
+    </Dialog>
+  );
+}
+
+// ---------- Sub-components ----------
+
+function ContextBadge({
+  status,
+  runtimeCount,
+}: {
+  status: DaemonStatus;
+  runtimeCount: number;
+}) {
+  const isRunning = status.state === "running";
+  return (
+    <span className="inline-flex items-center gap-1.5 rounded-md border bg-background px-1.5 py-0.5 text-xs font-normal">
+      <span
+        className={cn(
+          "size-1.5 rounded-full",
+          DAEMON_STATE_COLORS[status.state],
+        )}
+      />
+      <span
+        className={cn(
+          "tabular-nums",
+          isRunning ? "text-foreground" : "text-muted-foreground",
+        )}
+      >
+        {DAEMON_STATE_LABELS[status.state]}
+      </span>
+      {isRunning && status.uptime && (
+        <span className="text-muted-foreground">
+          · {formatUptime(status.uptime)}
+        </span>
+      )}
+      {isRunning && runtimeCount > 0 && (
+        <span className="text-muted-foreground">
+          · {runtimeCount} runtime{runtimeCount === 1 ? "" : "s"}
+        </span>
+      )}
+    </span>
+  );
+}
+
+function FilterChip({
+  active,
+  onClick,
+  label,
+  count,
+  variant,
+}: {
+  active: boolean;
+  onClick: () => void;
+  label: string;
+  count: number;
+  variant?: LogLevel;
+}) {
+  return (
+    <button
+      type="button"
+      onClick={onClick}
+      className={cn(
+        "inline-flex h-7 items-center gap-1 rounded-md border bg-background px-2 text-xs transition-colors hover:bg-accent",
+        active
+          ? variant
+            ? LEVEL_BADGE_CLASS[variant]
+            : "bg-accent text-accent-foreground"
+          : "border-dashed text-muted-foreground/50",
+      )}
+    >
+      {label}
+      <span
+        className={cn(
+          "tabular-nums",
+          active ? "text-current/80" : "text-muted-foreground/40",
+        )}
+      >
+        {count}
+      </span>
+    </button>
+  );
+}
+
+function LevelBadge({ level }: { level: LogLevel }) {
+  return (
+    <span
+      className={cn(
+        "inline-flex h-4 shrink-0 items-center rounded border px-1 text-[10px] font-medium uppercase tracking-wide",
+        LEVEL_BADGE_CLASS[level],
+      )}
+    >
+      {level}
+    </span>
+  );
+}
+
+function LogLineRow({
+  line,
+  expanded,
+  onToggle,
+  search,
+}: {
+  line: ParsedLogLine;
+  expanded: boolean;
+  onToggle: () => void;
+  search: string;
+}) {
+  const fieldEntries = Object.entries(line.fields);
+  const hasFields = fieldEntries.length > 0;
+
+  // Unparseable line — render the raw text so nothing is hidden. Common
+  // for panic stack traces and partial writes during log rotation.
+  if (!line.timestamp || !line.level) {
+    return (
+      <div className="break-all whitespace-pre-wrap px-2 py-0.5 text-muted-foreground/70">
+        {highlight(line.raw, search)}
+      </div>
+    );
+  }
+
+  return (
+    <div
+      className={cn(
+        "grid grid-cols-[auto_auto_minmax(0,1fr)] items-baseline gap-2 rounded px-2 py-0.5 hover:bg-accent/30",
+        hasFields && "cursor-pointer",
+      )}
+      onClick={hasFields ? onToggle : undefined}
+    >
+      <span className="shrink-0 tabular-nums text-muted-foreground/60">
+        {line.timestamp}
+      </span>
+      <LevelBadge level={line.level} />
+      <div className="min-w-0">
+        <div className="flex min-w-0 items-baseline gap-2">
+          <span className="break-words">{highlight(line.message, search)}</span>
+          {hasFields && !expanded && (
+            <span className="min-w-0 truncate text-muted-foreground/60">
+              {fieldEntries
+                .map(([k, v]) => `${k}=${truncateValue(v)}`)
+                .join("  ")}
+            </span>
+          )}
+        </div>
+        {expanded && hasFields && (
+          <div className="ml-1 mt-1 grid grid-cols-[max-content_minmax(0,1fr)] gap-x-3 gap-y-0.5 text-muted-foreground">
+            {fieldEntries.map(([k, v]) => (
+              <Fragment key={k}>
+                <span className="text-muted-foreground/70">{k}</span>
+                <span className="break-all text-foreground/85">{v}</span>
+              </Fragment>
+            ))}
+          </div>
+        )}
+      </div>
+    </div>
+  );
+}
+
+function GroupRows({
+  first,
+  rest,
+  expanded,
+  onToggle,
+  expandedFields,
+  onToggleFields,
+  search,
+}: {
+  first: ParsedLogLine;
+  rest: ParsedLogLine[];
+  expanded: boolean;
+  onToggle: () => void;
+  expandedFields: Set<number>;
+  onToggleFields: (id: number) => void;
+  search: string;
+}) {
+  // Folded: show the first occurrence so the user still sees a sample
+  // (timestamp, level, message), then a click-to-expand placeholder for
+  // the suppressed run. The placeholder uses a dashed border + italics
+  // so the eye reads it as "not a real line".
+  if (!expanded) {
+    return (
+      <>
+        <LogLineRow
+          line={first}
+          expanded={expandedFields.has(first.id)}
+          onToggle={() => onToggleFields(first.id)}
+          search={search}
+        />
+        <button
+          type="button"
+          onClick={onToggle}
+          className="my-0.5 ml-2 inline-flex w-fit items-center gap-2 rounded border border-dashed border-muted-foreground/25 bg-muted/30 px-2 py-0.5 text-[11px] italic text-muted-foreground/70 hover:bg-muted/60 hover:text-foreground"
+        >
+          <span>···</span>
+          <span>
+            {rest.length} more &ldquo;{truncateValue(first.message, 48)}
+            &rdquo; — click to expand
+          </span>
+        </button>
+      </>
+    );
+  }
+
+  // Unfolded: render every line, then a small "collapse" affordance at
+  // the end so the user can put the toothpaste back in the tube.
+  return (
+    <>
+      <LogLineRow
+        line={first}
+        expanded={expandedFields.has(first.id)}
+        onToggle={() => onToggleFields(first.id)}
+        search={search}
+      />
+      {rest.map((l) => (
+        <LogLineRow
+          key={l.id}
+          line={l}
+          expanded={expandedFields.has(l.id)}
+          onToggle={() => onToggleFields(l.id)}
+          search={search}
+        />
+      ))}
+      <button
+        type="button"
+        onClick={onToggle}
+        className="my-0.5 ml-2 inline-flex w-fit items-center gap-2 rounded border border-dashed border-muted-foreground/25 px-2 py-0.5 text-[11px] italic text-muted-foreground/60 hover:text-foreground"
+      >
+        <span>···</span>
+        <span>collapse {rest.length + 1} repeated</span>
+      </button>
+    </>
+  );
+}
+
+function EmptyState({
+  hasLogs,
+  hasFilter,
+  isRunning,
+}: {
+  hasLogs: boolean;
+  hasFilter: boolean;
+  isRunning: boolean;
+}) {
+  let title: string;
+  let subtitle: string;
+  if (hasFilter) {
+    title = "No matching log lines";
+    subtitle = "Try a different search or level toggle.";
+  } else if (!isRunning) {
+    title = "Daemon isn't running";
+    subtitle = "Start the daemon to see logs here.";
+  } else if (!hasLogs) {
+    title = "Waiting for logs…";
+    subtitle = "New entries will appear in real time.";
+  } else {
+    title = "";
+    subtitle = "";
+  }
+  return (
+    <div className="flex h-full flex-col items-center justify-center gap-1 text-center text-muted-foreground/70">
+      <p className="text-sm">{title}</p>
+      <p className="text-xs text-muted-foreground/50">{subtitle}</p>
+    </div>
+  );
+}
+
+// ---------- Helpers ----------
+
+function truncateValue(value: string, max = 32): string {
+  return value.length > max ? `${value.slice(0, max)}…` : value;
+}
+
+function highlight(text: string, query: string): ReactNode {
+  if (!query) return text;
+  const q = query.toLowerCase();
+  const lower = text.toLowerCase();
+  const idx = lower.indexOf(q);
+  if (idx === -1) return text;
+  return (
+    <>
+      {text.slice(0, idx)}
+      <mark className="rounded bg-warning/30 px-0.5 text-foreground">
+        {text.slice(idx, idx + query.length)}
+      </mark>
+      {text.slice(idx + query.length)}
+    </>
+  );
+}

apps/desktop/src/renderer/src/components/daemon-runtime-card.tsx

@@ -0,0 +1,332 @@
+import { useState, useEffect, useCallback, useMemo } from "react";
+import {
+  AlertCircle,
+  Play,
+  Square,
+  RotateCw,
+  Server,
+  Activity,
+  ScrollText,
+} from "lucide-react";
+import { useQuery } from "@tanstack/react-query";
+import { useWorkspaceId } from "@multica/core/hooks";
+import { runtimeListOptions } from "@multica/core/runtimes";
+import { agentTaskSnapshotOptions } from "@multica/core/agents";
+import { cn } from "@multica/ui/lib/utils";
+import { Button } from "@multica/ui/components/ui/button";
+import {
+  Card,
+  CardAction,
+  CardDescription,
+  CardHeader,
+  CardTitle,
+} from "@multica/ui/components/ui/card";
+import {
+  Dialog,
+  DialogContent,
+  DialogDescription,
+  DialogFooter,
+  DialogHeader,
+  DialogTitle,
+} from "@multica/ui/components/ui/dialog";
+import { toast } from "sonner";
+import { DaemonPanel } from "./daemon-panel";
+import type { DaemonStatus } from "../../../shared/daemon-types";
+import {
+  DAEMON_STATE_COLORS,
+  DAEMON_STATE_LABELS,
+  daemonStateDescription,
+  formatUptime,
+} from "../../../shared/daemon-types";
+
+/**
+ * Header card on the desktop Runtimes page that surfaces the daemon embedded
+ * in this Electron app. The same daemon process registers N runtimes with the
+ * server (one per detected CLI), which appear in the runtime list below — so
+ * this card is the parent control surface for "what's running on this Mac".
+ *
+ * Why this lives only on desktop: web users don't have an embedded daemon;
+ * they bring their own (CLI-launched or remote VM) and just see runtimes in
+ * the list. The `desktop-runtimes-page` wrapper is the only mount point.
+ */
+export function DaemonRuntimeCard() {
+  const [status, setStatus] = useState<DaemonStatus>({ state: "stopped" });
+  const [panelOpen, setPanelOpen] = useState(false);
+  const [actionLoading, setActionLoading] = useState(false);
+  const [confirmStop, setConfirmStop] = useState(false);
+
+  const wsId = useWorkspaceId();
+  const { data: runtimes = [] } = useQuery(runtimeListOptions(wsId));
+  // Snapshot also includes each agent's latest terminal; the filter below
+  // drops anything that isn't running/dispatched, so terminal rows pass
+  // through harmlessly.
+  const { data: snapshot = [] } = useQuery(agentTaskSnapshotOptions(wsId));
+
+  // Set of runtime IDs registered by THIS daemon (one per detected CLI).
+  // Used both to count "how many CLIs am I contributing" and to figure
+  // out which active tasks would be impacted by a Stop.
+  const localRuntimeIds = useMemo(() => {
+    if (!status.daemonId) return new Set<string>();
+    return new Set(
+      runtimes
+        .filter((r) => r.daemon_id === status.daemonId)
+        .map((r) => r.id),
+    );
+  }, [runtimes, status.daemonId]);
+
+  const runtimeCount = localRuntimeIds.size;
+
+  // Tasks that are actually doing work on this daemon right now —
+  // running or dispatched. Queued tasks haven't claimed a runtime yet,
+  // so stopping the daemon won't break them (they'll wait for any
+  // available daemon). The number drives the Stop-confirmation dialog.
+  const affectedTasks = useMemo(
+    () =>
+      snapshot.filter(
+        (t) =>
+          localRuntimeIds.has(t.runtime_id) &&
+          (t.status === "running" || t.status === "dispatched"),
+      ),
+    [snapshot, localRuntimeIds],
+  );
+
+  useEffect(() => {
+    window.daemonAPI.getStatus().then((s) => setStatus(s));
+    const unsub = window.daemonAPI.onStatusChange((s) => {
+      setStatus(s);
+      setActionLoading(false);
+    });
+    return unsub;
+  }, []);
+
+  const handleStart = useCallback(async () => {
+    setActionLoading(true);
+    const result = await window.daemonAPI.start();
+    if (!result.success) {
+      setActionLoading(false);
+      toast.error("Failed to start daemon", { description: result.error });
+    }
+  }, []);
+
+  // The actual stop call, separated from the click handler so we can call
+  // it both from the direct path (no active tasks) and from the confirm
+  // dialog's confirm button.
+  const performStop = useCallback(async () => {
+    setActionLoading(true);
+    const result = await window.daemonAPI.stop();
+    if (!result.success) {
+      toast.error("Failed to stop daemon", { description: result.error });
+    }
+  }, []);
+
+  // Click on the Stop button. If there's nothing running, just stop;
+  // otherwise pop a confirm dialog explaining the blast radius.
+  const handleStopClick = useCallback(() => {
+    if (affectedTasks.length === 0) {
+      void performStop();
+    } else {
+      setConfirmStop(true);
+    }
+  }, [affectedTasks.length, performStop]);
+
+  const handleRestart = useCallback(async () => {
+    setActionLoading(true);
+    const result = await window.daemonAPI.restart();
+    if (!result.success) {
+      toast.error("Failed to restart daemon", { description: result.error });
+      return;
+    }
+    // Success feedback — the daemon takes a few seconds to come back online,
+    // and the only other UI signal is the state badge flipping briefly. A
+    // toast confirms the click was received and tells the user what to expect.
+    toast.success("Restarting daemon", {
+      description: "Runtimes will be back online in a few seconds.",
+    });
+  }, []);
+
+  const handleRetryInstall = useCallback(async () => {
+    setActionLoading(true);
+    try {
+      await window.daemonAPI.retryInstall();
+    } finally {
+      setActionLoading(false);
+    }
+  }, []);
+
+  const isRunning = status.state === "running";
+  const isStopped = status.state === "stopped";
+  const isCliMissing = status.state === "cli_not_found";
+  const isTransitioning =
+    status.state === "starting" || status.state === "stopping";
+  const isInstalling = status.state === "installing_cli";
+
+  return (
+    <>
+      <Card size="sm">
+        <CardHeader>
+          <CardTitle className="flex items-center gap-2">
+            <Server className="size-4 text-muted-foreground" />
+            Local daemon
+            <span className="inline-flex items-center gap-1.5 rounded-md border bg-background px-1.5 py-0.5 text-xs font-normal">
+              <span
+                className={cn(
+                  "size-1.5 rounded-full",
+                  DAEMON_STATE_COLORS[status.state],
+                )}
+              />
+              <span
+                className={cn(
+                  "tabular-nums",
+                  isRunning ? "text-foreground" : "text-muted-foreground",
+                )}
+              >
+                {DAEMON_STATE_LABELS[status.state]}
+              </span>
+              {isRunning && status.uptime && (
+                <span className="text-muted-foreground">
+                  · {formatUptime(status.uptime)}
+                </span>
+              )}
+            </span>
+          </CardTitle>
+          <CardDescription>
+            {daemonStateDescription(status.state, runtimeCount)}
+          </CardDescription>
+          <CardAction className="self-center">
+            <div className="flex items-center gap-1.5">
+              {isRunning && (
+                <>
+                  <Button
+                    size="sm"
+                    variant="ghost"
+                    onClick={() => setPanelOpen(true)}
+                  >
+                    <ScrollText className="size-3.5 mr-1.5" />
+                    View logs
+                  </Button>
+                  <Button
+                    size="sm"
+                    variant="outline"
+                    onClick={handleRestart}
+                    disabled={actionLoading}
+                  >
+                    <RotateCw className="size-3.5 mr-1.5" />
+                    Restart
+                  </Button>
+                  <Button
+                    size="sm"
+                    variant="destructive"
+                    onClick={handleStopClick}
+                    disabled={actionLoading}
+                  >
+                    <Square className="size-3.5 mr-1.5" />
+                    Stop
+                  </Button>
+                </>
+              )}
+
+              {isStopped && (
+                <Button
+                  size="sm"
+                  onClick={handleStart}
+                  disabled={actionLoading}
+                >
+                  {actionLoading ? (
+                    <Activity className="size-3.5 mr-1.5 animate-pulse" />
+                  ) : (
+                    <Play className="size-3.5 mr-1.5" />
+                  )}
+                  Start
+                </Button>
+              )}
+
+              {isCliMissing && (
+                <Button
+                  size="sm"
+                  variant="outline"
+                  onClick={handleRetryInstall}
+                  disabled={actionLoading}
+                >
+                  <RotateCw className="size-3.5 mr-1.5" />
+                  Retry setup
+                </Button>
+              )}
+
+              {(isTransitioning || isInstalling) && (
+                <Button size="sm" variant="outline" disabled>
+                  <Activity className="size-3.5 mr-1.5 animate-pulse" />
+                  {DAEMON_STATE_LABELS[status.state]}
+                </Button>
+              )}
+            </div>
+          </CardAction>
+        </CardHeader>
+      </Card>
+
+      <DaemonPanel
+        open={panelOpen}
+        onOpenChange={setPanelOpen}
+        status={status}
+        runtimeCount={runtimeCount}
+      />
+
+      <StopConfirmDialog
+        open={confirmStop}
+        onOpenChange={setConfirmStop}
+        affectedCount={affectedTasks.length}
+        onConfirm={() => {
+          setConfirmStop(false);
+          void performStop();
+        }}
+      />
+    </>
+  );
+}
+
+// ---------- Sub-components ----------
+
+function StopConfirmDialog({
+  open,
+  onOpenChange,
+  affectedCount,
+  onConfirm,
+}: {
+  open: boolean;
+  onOpenChange: (v: boolean) => void;
+  affectedCount: number;
+  onConfirm: () => void;
+}) {
+  const plural = affectedCount === 1 ? "" : "s";
+  const verb = affectedCount === 1 ? "is" : "are";
+
+  return (
+    <Dialog open={open} onOpenChange={onOpenChange}>
+      <DialogContent className="max-w-sm" showCloseButton={false}>
+        <div className="flex items-start gap-3">
+          <div className="flex h-10 w-10 shrink-0 items-center justify-center rounded-full bg-destructive/10">
+            <AlertCircle className="h-5 w-5 text-destructive" />
+          </div>
+          <DialogHeader className="flex-1 gap-1">
+            <DialogTitle className="text-sm font-semibold">
+              Stop daemon with {affectedCount} active task{plural}?
+            </DialogTitle>
+            <DialogDescription className="text-xs leading-relaxed">
+              {affectedCount} task{plural} {verb} currently running on this
+              device. Stopping now will interrupt {affectedCount === 1 ? "it" : "them"}{" "}
+              — affected tasks get marked <strong>failed</strong> once the
+              timeout hits. The daemon won&apos;t auto-restart.
+            </DialogDescription>
+          </DialogHeader>
+        </div>
+        <DialogFooter>
+          <Button variant="ghost" onClick={() => onOpenChange(false)}>
+            Cancel
+          </Button>
+          <Button variant="destructive" onClick={onConfirm}>
+            Stop daemon
+          </Button>
+        </DialogFooter>
+      </DialogContent>
+    </Dialog>
+  );
+}

apps/desktop/src/renderer/src/components/daemon-settings-tab.tsx

@@ -0,0 +1,201 @@
+import { useState, useEffect, useCallback, type ReactNode } from "react";
+import { Button } from "@multica/ui/components/ui/button";
+import { Switch } from "@multica/ui/components/ui/switch";
+import { cn } from "@multica/ui/lib/utils";
+import type { DaemonPrefs, DaemonStatus } from "../../../shared/daemon-types";
+import {
+  DAEMON_STATE_COLORS,
+  DAEMON_STATE_LABELS,
+  formatUptime,
+} from "../../../shared/daemon-types";
+
+function SettingRow({
+  label,
+  description,
+  children,
+}: {
+  label: string;
+  description: string;
+  children: ReactNode;
+}) {
+  return (
+    <div className="flex items-center justify-between gap-6 py-4">
+      <div className="min-w-0">
+        <p className="text-sm font-medium">{label}</p>
+        <p className="text-sm text-muted-foreground mt-0.5">{description}</p>
+      </div>
+      <div className="shrink-0">{children}</div>
+    </div>
+  );
+}
+
+// One row inside the diagnostics block. Values that are likely to be
+// long IDs / URLs render as monospaced + truncated with a tooltip.
+function DiagnosticsRow({
+  label,
+  value,
+  mono,
+}: {
+  label: string;
+  value: ReactNode;
+  mono?: boolean;
+}) {
+  return (
+    <div className="grid grid-cols-[140px_minmax(0,1fr)] items-baseline gap-3 py-1.5">
+      <span className="text-xs text-muted-foreground">{label}</span>
+      <span
+        className={cn(
+          "min-w-0 truncate text-sm",
+          mono && "font-mono text-xs",
+        )}
+        title={typeof value === "string" ? value : undefined}
+      >
+        {value}
+      </span>
+    </div>
+  );
+}
+
+export function DaemonSettingsTab() {
+  const [prefs, setPrefs] = useState<DaemonPrefs>({ autoStart: true, autoStop: false });
+  const [cliInstalled, setCliInstalled] = useState<boolean | null>(null);
+  const [saving, setSaving] = useState(false);
+  const [status, setStatus] = useState<DaemonStatus>({ state: "stopped" });
+
+  useEffect(() => {
+    window.daemonAPI.getPrefs().then(setPrefs);
+    window.daemonAPI.isCliInstalled().then(setCliInstalled);
+    window.daemonAPI.getStatus().then(setStatus);
+    return window.daemonAPI.onStatusChange(setStatus);
+  }, []);
+
+  const updatePref = useCallback(
+    async (key: keyof DaemonPrefs, value: boolean) => {
+      setSaving(true);
+      const updated = await window.daemonAPI.setPrefs({ [key]: value });
+      setPrefs(updated);
+      setSaving(false);
+    },
+    [],
+  );
+
+  return (
+    <div>
+      <h2 className="text-lg font-semibold">Daemon</h2>
+      <p className="text-sm text-muted-foreground mt-1">
+        Configure how the local agent daemon behaves with the desktop app.
+      </p>
+
+      <div className="mt-6 divide-y">
+        <SettingRow
+          label="Auto-start on launch"
+          description="Automatically start the daemon when the app opens and you are logged in."
+        >
+          <Switch
+            checked={prefs.autoStart}
+            onCheckedChange={(checked) => updatePref("autoStart", checked)}
+            disabled={saving}
+          />
+        </SettingRow>
+
+        <SettingRow
+          label="Auto-stop on quit"
+          description="Stop the daemon when the desktop app is closed. Disable this to keep the daemon running in the background."
+        >
+          <Switch
+            checked={prefs.autoStop}
+            onCheckedChange={(checked) => updatePref("autoStop", checked)}
+            disabled={saving}
+          />
+        </SettingRow>
+
+        <div className="py-4">
+          <p className="text-sm font-medium">CLI Status</p>
+          <p className="text-sm text-muted-foreground mt-1">
+            {cliInstalled === null
+              ? "Checking…"
+              : cliInstalled
+                ? "multica CLI is installed and available in PATH."
+                : "multica CLI not found. Install it to enable daemon management."}
+          </p>
+          {cliInstalled === false && (
+            <Button
+              variant="outline"
+              size="sm"
+              className="mt-2"
+              onClick={() =>
+                window.desktopAPI.openExternal(
+                  "https://github.com/multica-ai/multica#cli-installation",
+                )
+              }
+            >
+              Installation Guide
+            </Button>
+          )}
+        </div>
+      </div>
+
+      {/* Diagnostics — moved out of the logs panel so the panel can focus
+          on logs. These fields matter for support tickets and bug reports,
+          not for everyday use. */}
+      <div className="mt-8">
+        <h3 className="text-sm font-semibold">Diagnostics</h3>
+        <p className="text-xs text-muted-foreground mt-1">
+          Identification and connection details. Useful when filing a bug
+          report or investigating why a runtime isn&apos;t showing up.
+        </p>
+        <div className="mt-3 rounded-lg border bg-muted/20 px-4 py-2">
+          <DiagnosticsRow
+            label="State"
+            value={
+              <span className="inline-flex items-center gap-1.5">
+                <span
+                  className={cn(
+                    "size-1.5 rounded-full",
+                    DAEMON_STATE_COLORS[status.state],
+                  )}
+                />
+                {DAEMON_STATE_LABELS[status.state]}
+              </span>
+            }
+          />
+          <DiagnosticsRow
+            label="Uptime"
+            value={status.uptime ? formatUptime(status.uptime) : "—"}
+          />
+          <DiagnosticsRow
+            label="PID"
+            value={status.pid ?? "—"}
+            mono={!!status.pid}
+          />
+          <DiagnosticsRow
+            label="Daemon ID"
+            value={status.daemonId ?? "—"}
+            mono={!!status.daemonId}
+          />
+          <DiagnosticsRow
+            label="Profile"
+            value={status.profile || "default"}
+          />
+          <DiagnosticsRow
+            label="Server URL"
+            value={status.serverUrl ?? "—"}
+            mono={!!status.serverUrl}
+          />
+          <DiagnosticsRow
+            label="Device name"
+            value={status.deviceName ?? "—"}
+          />
+          <DiagnosticsRow
+            label="Workspaces"
+            value={
+              typeof status.workspaceCount === "number"
+                ? status.workspaceCount
+                : "—"
+            }
+          />
+        </div>
+      </div>
+    </div>
+  );
+}

apps/desktop/src/renderer/src/components/desktop-layout.tsx

@@ -0,0 +1,177 @@
+import { useEffect, useSyncExternalStore } from "react";
+import { ChevronLeft, ChevronRight } from "lucide-react";
+import { cn } from "@multica/ui/lib/utils";
+import { useTabHistory } from "@/hooks/use-tab-history";
+import { useActiveTitleSync } from "@/hooks/use-tab-sync";
+import { useTabStore, resolveRouteIcon } from "@/stores/tab-store";
+import {
+  SidebarProvider,
+  SidebarTrigger,
+  useSidebar,
+} from "@multica/ui/components/ui/sidebar";
+import { ModalRegistry } from "@multica/views/modals/registry";
+import { AppSidebar } from "@multica/views/layout";
+import { SearchCommand, SearchTrigger } from "@multica/views/search";
+import { ChatFab, ChatWindow } from "@multica/views/chat";
+import { StarterContentPrompt } from "@multica/views/onboarding";
+import { WorkspaceSlugProvider, paths, useCurrentWorkspace } from "@multica/core/paths";
+import { getCurrentSlug, subscribeToCurrentSlug } from "@multica/core/platform";
+import { useDesktopUnreadBadge } from "@multica/views/platform";
+import { DesktopNavigationProvider } from "@/platform/navigation";
+import { TabBar } from "./tab-bar";
+import { TabContent } from "./tab-content";
+import { WindowOverlay } from "./window-overlay";
+
+function SidebarTopBar() {
+  const { canGoBack, canGoForward, goBack, goForward } = useTabHistory();
+
+  return (
+    <div
+      className="h-12 shrink-0 flex items-center justify-end px-2"
+      style={{ WebkitAppRegion: "drag" } as React.CSSProperties}
+    >
+      <div
+        className="flex items-center gap-0.5"
+        style={{ WebkitAppRegion: "no-drag" } as React.CSSProperties}
+      >
+        <button
+          onClick={goBack}
+          disabled={!canGoBack}
+          aria-label="Go back"
+          className="flex size-7 items-center justify-center rounded-md text-muted-foreground transition-colors hover:bg-accent hover:text-foreground disabled:opacity-30 disabled:pointer-events-none"
+        >
+          <ChevronLeft className="size-4" />
+        </button>
+        <button
+          onClick={goForward}
+          disabled={!canGoForward}
+          aria-label="Go forward"
+          className="flex size-7 items-center justify-center rounded-md text-muted-foreground transition-colors hover:bg-accent hover:text-foreground disabled:opacity-30 disabled:pointer-events-none"
+        >
+          <ChevronRight className="size-4" />
+        </button>
+      </div>
+    </div>
+  );
+}
+
+// The main area's top bar doubles as a window drag region. When the sidebar
+// is not occupying main-flow width — either user-collapsed (offcanvas) or
+// auto-hidden in mobile mode (<768px, becomes a sheet drawer) — we pad the
+// left side so tabs don't land under the macOS traffic lights (which live at
+// roughly x=16..68 and always hit-test above HTML), and surface a trigger so
+// the sidebar can be brought back without keyboard shortcut.
+function MainTopBar() {
+  const { state, isMobile } = useSidebar();
+  const sidebarHidden = state === "collapsed" || isMobile;
+
+  return (
+    <header
+      className={cn(
+        "h-12 shrink-0 flex items-center gap-2",
+        sidebarHidden && "pl-20",
+      )}
+      style={{ WebkitAppRegion: "drag" } as React.CSSProperties}
+    >
+      {sidebarHidden && (
+        <SidebarTrigger
+          style={{ WebkitAppRegion: "no-drag" } as React.CSSProperties}
+        />
+      )}
+      <TabBar />
+    </header>
+  );
+}
+
+function useInternalLinkHandler() {
+  useEffect(() => {
+    const handler = (e: Event) => {
+      const path = (e as CustomEvent).detail?.path;
+      if (!path) return;
+      const icon = resolveRouteIcon(path);
+      const store = useTabStore.getState();
+      const tabId = store.openTab(path, path, icon);
+      store.setActiveTab(tabId);
+    };
+    window.addEventListener("multica:navigate", handler);
+    return () => window.removeEventListener("multica:navigate", handler);
+  }, []);
+}
+
+/**
+ * Bridge between the renderer and the Electron main process for inbox-level
+ * OS integration. Mounted inside WorkspaceSlugProvider so it can resolve the
+ * current workspace's id for the badge hook.
+ *
+ * Two responsibilities:
+ *   1. Mirror the unread inbox count onto the dock/taskbar badge.
+ *   2. When the user clicks an OS notification, open the notified
+ *      workspace's inbox focused on that item. The route uses the `slug`
+ *      that the notification was *emitted* with — not the currently active
+ *      workspace — so a notification from workspace A always opens A's
+ *      inbox even if the user has since switched to workspace B. Marking
+ *      the row read is handled by InboxPage's selected-item effect, which
+ *      covers both click-to-select and URL-param-select paths.
+ */
+function DesktopInboxBridge() {
+  const workspace = useCurrentWorkspace();
+  useDesktopUnreadBadge(workspace?.id ?? null);
+
+  useEffect(() => {
+    return window.desktopAPI.onInboxOpen(({ slug, issueKey }) => {
+      if (!slug) return;
+      const inboxPath = `${paths.workspace(slug).inbox()}?issue=${encodeURIComponent(issueKey)}`;
+      window.dispatchEvent(
+        new CustomEvent("multica:navigate", { detail: { path: inboxPath } }),
+      );
+    });
+  }, []);
+
+  return null;
+}
+
+export function DesktopShell() {
+  useInternalLinkHandler();
+  useActiveTitleSync();
+
+  // Reactive read of current workspace slug from the platform singleton.
+  // On first mount, slug is null until WorkspaceRouteLayout (inside the tab
+  // router) sets it. Once set, the sidebar and other shell-level components
+  // can resolve workspace-scoped paths via useWorkspacePaths().
+  const slug = useSyncExternalStore(subscribeToCurrentSlug, getCurrentSlug, () => null);
+
+  return (
+    <DesktopNavigationProvider>
+      {/* WorkspaceSlugProvider accepts null — components that need slug
+          use useWorkspaceSlug() (nullable) or useRequiredWorkspaceSlug()
+          (throws). TabContent MUST always render so the tab router can
+          mount WorkspaceRouteLayout, which calls setCurrentWorkspace()
+          to populate the slug. The sidebar gates on slug being present
+          to avoid the useRequiredWorkspaceSlug throw. Zero-workspace
+          users see the window-level overlay (new-workspace flow)
+          triggered by IndexRedirect, not a route. */}
+      <WorkspaceSlugProvider slug={slug}>
+        <DesktopInboxBridge />
+        <div className="flex h-screen">
+          <SidebarProvider className="flex-1">
+            {slug && <AppSidebar topSlot={<SidebarTopBar />} searchSlot={<SearchTrigger />} />}
+            {/* Right side: header + content container */}
+            <div className="flex flex-1 min-w-0 flex-col">
+              <MainTopBar />
+              {/* Content area with inset styling — relative so ChatWindow/ChatFab are constrained here */}
+              <div className="relative flex flex-1 min-h-0 flex-col overflow-hidden mr-2 mb-2 ml-0.5 rounded-xl shadow-sm bg-background">
+                <TabContent />
+                {slug && <ChatWindow />}
+                {slug && <ChatFab />}
+              </div>
+            </div>
+          </SidebarProvider>
+        </div>
+        {slug && <ModalRegistry />}
+        {slug && <SearchCommand />}
+        {slug && <StarterContentPrompt />}
+        <WindowOverlay />
+      </WorkspaceSlugProvider>
+    </DesktopNavigationProvider>
+  );
+}

apps/desktop/src/renderer/src/components/desktop-runtimes-page.tsx

@@ -0,0 +1,39 @@
+import { useEffect, useState } from "react";
+import { RuntimesPage } from "@multica/views/runtimes";
+import { DaemonRuntimeCard } from "./daemon-runtime-card";
+import type { DaemonStatus } from "../../../shared/daemon-types";
+
+/**
+ * Desktop wrapper around the shared `RuntimesPage`. Bridges the Electron
+ * `daemonAPI` (main-process daemon state) into the page so its empty
+ * state can distinguish "no runtime registered" from "runtime is on its
+ * way" — without the bundled daemon's status, the page shows a
+ * misleading "Run multica daemon start" hint during the few seconds
+ * between page load and the daemon's first registration.
+ *
+ * `bootstrapping` is true while the daemon is installing, starting, or
+ * already running but hasn't surfaced as a server-side runtime yet.
+ * RuntimeList only shows the spinner when the runtime list is also
+ * empty, so once the daemon registers (and the list fills) the flag
+ * has no visible effect.
+ */
+export function DesktopRuntimesPage() {
+  const [status, setStatus] = useState<DaemonStatus>({ state: "stopped" });
+
+  useEffect(() => {
+    window.daemonAPI.getStatus().then(setStatus);
+    return window.daemonAPI.onStatusChange(setStatus);
+  }, []);
+
+  const bootstrapping =
+    status.state === "installing_cli" ||
+    status.state === "starting" ||
+    status.state === "running";
+
+  return (
+    <RuntimesPage
+      topSlot={<DaemonRuntimeCard />}
+      bootstrapping={bootstrapping}
+    />
+  );
+}

apps/desktop/src/renderer/src/components/pageview-tracker.test.tsx

@@ -0,0 +1,243 @@
+import { describe, expect, it, vi, beforeEach } from "vitest";
+import { render } from "@testing-library/react";
+
+// vi.hoisted shared state — every store mock reads the same object so each
+// test can mutate it then re-render to drive the tracker.
+const state = vi.hoisted(() => ({
+  user: null as { id: string } | null,
+  overlay: null as { type: string; invitationId?: string } | null,
+  activeWorkspaceSlug: null as string | null,
+  byWorkspace: {} as Record<
+    string,
+    { activeTabId: string; tabs: { id: string; path: string }[] }
+  >,
+  capturePageview: vi.fn<(path?: string) => void>(),
+}));
+
+vi.mock("@multica/core/analytics", () => ({
+  capturePageview: state.capturePageview,
+}));
+
+// Auth store — single selector pattern (`s => s.user`).
+vi.mock("@multica/core/auth", () => {
+  const useAuthStore = (selector: (s: typeof state) => unknown) =>
+    selector(state);
+  return { useAuthStore };
+});
+
+// Window overlay store — same shape.
+vi.mock("@/stores/window-overlay-store", () => {
+  const useWindowOverlayStore = (selector: (s: typeof state) => unknown) =>
+    selector(state);
+  return { useWindowOverlayStore };
+});
+
+// Tab store — selectors read activeWorkspaceSlug + byWorkspace. Also expose
+// getState() for the seed pass and the helpers the tracker imports
+// (useActiveTabIdentity, getActiveTab) so we don't have to re-import them
+// from the real store inside a mocked module.
+vi.mock("@/stores/tab-store", () => {
+  const useTabStore = Object.assign(
+    (selector: (s: typeof state) => unknown) => selector(state),
+    { getState: () => state },
+  );
+  const getActiveTab = (s: typeof state) => {
+    const slug = s.activeWorkspaceSlug;
+    if (!slug) return null;
+    const group = s.byWorkspace[slug];
+    if (!group) return null;
+    return group.tabs.find((t) => t.id === group.activeTabId) ?? null;
+  };
+  const useActiveTabIdentity = () => ({
+    slug: state.activeWorkspaceSlug,
+    tabId: state.activeWorkspaceSlug
+      ? (state.byWorkspace[state.activeWorkspaceSlug]?.activeTabId ?? null)
+      : null,
+  });
+  return { useTabStore, getActiveTab, useActiveTabIdentity };
+});
+
+import { PageviewTracker } from "./pageview-tracker";
+
+function reset() {
+  state.user = { id: "u1" };
+  state.overlay = null;
+  state.activeWorkspaceSlug = null;
+  state.byWorkspace = {};
+  state.capturePageview.mockClear();
+}
+
+beforeEach(() => {
+  reset();
+});
+
+describe("PageviewTracker", () => {
+  it("suppresses pageview when switching to a previously-visible tab on its existing path", () => {
+    state.byWorkspace = {
+      acme: {
+        activeTabId: "tA",
+        tabs: [
+          { id: "tA", path: "/acme/issues" },
+          { id: "tB", path: "/acme/inbox" },
+        ],
+      },
+    };
+    state.activeWorkspaceSlug = "acme";
+
+    const { rerender } = render(<PageviewTracker />);
+    // Initial mount on tA — seeded as observed, no pageview because both
+    // tabs were already in the persisted store before the tracker mounted.
+    expect(state.capturePageview).not.toHaveBeenCalled();
+
+    // Switch to tB (already-known tab on its already-known path).
+    state.byWorkspace = {
+      acme: {
+        activeTabId: "tB",
+        tabs: [
+          { id: "tA", path: "/acme/issues" },
+          { id: "tB", path: "/acme/inbox" },
+        ],
+      },
+    };
+    rerender(<PageviewTracker />);
+    expect(state.capturePageview).not.toHaveBeenCalled();
+
+    // Switch back to tA — still no pageview.
+    state.byWorkspace = {
+      acme: {
+        activeTabId: "tA",
+        tabs: [
+          { id: "tA", path: "/acme/issues" },
+          { id: "tB", path: "/acme/inbox" },
+        ],
+      },
+    };
+    rerender(<PageviewTracker />);
+    expect(state.capturePageview).not.toHaveBeenCalled();
+  });
+
+  it("fires pageview when a new tab is opened (openInNewTab / addTab)", () => {
+    state.byWorkspace = {
+      acme: {
+        activeTabId: "tA",
+        tabs: [{ id: "tA", path: "/acme/issues" }],
+      },
+    };
+    state.activeWorkspaceSlug = "acme";
+
+    const { rerender } = render(<PageviewTracker />);
+    state.capturePageview.mockClear();
+
+    // Simulate openInNewTab("/acme/agents") → new tab tC added and activated.
+    state.byWorkspace = {
+      acme: {
+        activeTabId: "tC",
+        tabs: [
+          { id: "tA", path: "/acme/issues" },
+          { id: "tC", path: "/acme/agents" },
+        ],
+      },
+    };
+    rerender(<PageviewTracker />);
+
+    expect(state.capturePageview).toHaveBeenCalledTimes(1);
+    expect(state.capturePageview).toHaveBeenCalledWith("/acme/agents");
+  });
+
+  it("fires pageview when switchWorkspace opens a new path in another workspace", () => {
+    state.byWorkspace = {
+      acme: {
+        activeTabId: "tA",
+        tabs: [{ id: "tA", path: "/acme/issues" }],
+      },
+    };
+    state.activeWorkspaceSlug = "acme";
+
+    const { rerender } = render(<PageviewTracker />);
+    state.capturePageview.mockClear();
+
+    // Cross-workspace navigation: switchWorkspace("butter", "/butter/inbox")
+    // creates a fresh tab in the destination workspace and makes it active.
+    state.byWorkspace = {
+      acme: { activeTabId: "tA", tabs: [{ id: "tA", path: "/acme/issues" }] },
+      butter: {
+        activeTabId: "tD",
+        tabs: [{ id: "tD", path: "/butter/inbox" }],
+      },
+    };
+    state.activeWorkspaceSlug = "butter";
+    rerender(<PageviewTracker />);
+
+    expect(state.capturePageview).toHaveBeenCalledTimes(1);
+    expect(state.capturePageview).toHaveBeenCalledWith("/butter/inbox");
+  });
+
+  it("fires pageview on intra-tab navigation (path changes for the same tabId)", () => {
+    state.byWorkspace = {
+      acme: {
+        activeTabId: "tA",
+        tabs: [{ id: "tA", path: "/acme/issues" }],
+      },
+    };
+    state.activeWorkspaceSlug = "acme";
+
+    const { rerender } = render(<PageviewTracker />);
+    state.capturePageview.mockClear();
+
+    state.byWorkspace = {
+      acme: {
+        activeTabId: "tA",
+        tabs: [{ id: "tA", path: "/acme/issues/123" }],
+      },
+    };
+    rerender(<PageviewTracker />);
+
+    expect(state.capturePageview).toHaveBeenCalledTimes(1);
+    expect(state.capturePageview).toHaveBeenCalledWith("/acme/issues/123");
+  });
+
+  it("fires overlay and login pageviews and suppresses re-entry into the same tab afterward", () => {
+    state.byWorkspace = {
+      acme: {
+        activeTabId: "tA",
+        tabs: [{ id: "tA", path: "/acme/issues" }],
+      },
+    };
+    state.activeWorkspaceSlug = "acme";
+
+    const { rerender } = render(<PageviewTracker />);
+    state.capturePageview.mockClear();
+
+    // Open onboarding overlay.
+    state.overlay = { type: "onboarding" };
+    rerender(<PageviewTracker />);
+    expect(state.capturePageview).toHaveBeenLastCalledWith("/onboarding");
+
+    // Close overlay back to the tab — the tab is already observed on
+    // /acme/issues so this is a re-activation, no pageview.
+    state.capturePageview.mockClear();
+    state.overlay = null;
+    rerender(<PageviewTracker />);
+    expect(state.capturePageview).not.toHaveBeenCalled();
+
+    // Logout fires /login.
+    state.user = null;
+    rerender(<PageviewTracker />);
+    expect(state.capturePageview).toHaveBeenLastCalledWith("/login");
+  });
+
+  it("suppresses on initial mount when the active tab was restored from persistence", () => {
+    state.byWorkspace = {
+      acme: {
+        activeTabId: "tA",
+        tabs: [{ id: "tA", path: "/acme/issues" }],
+      },
+    };
+    state.activeWorkspaceSlug = "acme";
+
+    render(<PageviewTracker />);
+    // Restored tab — seeded, treated as a re-activation.
+    expect(state.capturePageview).not.toHaveBeenCalled();
+  });
+});
+

apps/desktop/src/renderer/src/components/pageview-tracker.tsx

@@ -0,0 +1,126 @@
+import { useEffect, useRef } from "react";
+import { capturePageview } from "@multica/core/analytics";
+import { useAuthStore } from "@multica/core/auth";
+import {
+  getActiveTab,
+  useActiveTabIdentity,
+  useTabStore,
+} from "@/stores/tab-store";
+import { useWindowOverlayStore, type WindowOverlay } from "@/stores/window-overlay-store";
+
+/**
+ * Fires a PostHog $pageview whenever the user's visible surface changes,
+ * EXCEPT for re-activations of an already-known tab on its already-known
+ * path.
+ *
+ * Desktop has three layers that can own the visible page:
+ *
+ *   1. Logged-out state → `/login`. No workspace context, no tabs.
+ *   2. Window overlays (onboarding, new-workspace, invite) → synthetic paths
+ *      that match the equivalent web routes. Overlays are NOT tab routes on
+ *      desktop (see `stores/window-overlay-store.ts` + `routes.tsx`), so the
+ *      tab path alone would either miss them or mislabel them as "/".
+ *   3. Otherwise → the active tab's path (workspace-scoped, e.g.
+ *      `/acme/issues/123`). Kept in sync by `useTabRouterSync`.
+ *
+ * Tab-switch suppression: re-activating an already-open tab surfaces a
+ * previously-visited path under a `(workspace, tabId)` we have already
+ * seen — the pageview was emitted when the user originally navigated
+ * there, so re-emitting on every switch just inflates PostHog billing
+ * without adding signal (real-data audit: desktop tab switches were
+ * ~50% of all `$pageview` events).
+ *
+ * Newly opened tabs (`openInNewTab`, `addTab`) and cross-workspace
+ * `switchWorkspace(slug, path)` to a previously-unseen tab still fire,
+ * because their key is not in the observed map yet. The map is seeded
+ * from the persisted tab store on first render so tabs restored from a
+ * previous session don't all re-emit on first activation.
+ *
+ * PostHog's `capture_pageview: true` auto-capture is intentionally off (see
+ * `initAnalytics`) so this component owns the event shape, matching the web
+ * implementation in `apps/web/components/pageview-tracker.tsx`.
+ */
+export function PageviewTracker() {
+  const user = useAuthStore((s) => s.user);
+  const overlay = useWindowOverlayStore((s) => s.overlay);
+  const { slug: activeWorkspaceSlug, tabId: activeTabId } = useActiveTabIdentity();
+  const activeTabPath = useTabStore((s) => getActiveTab(s)?.path ?? null);
+
+  // (slug:tabId) → last path observed while that tab was visible. Lets us
+  // tell "re-activating a tab on a path we already saw" (suppress) apart
+  // from "newly opened tab" or "intra-tab navigation" (fire). Seeded
+  // synchronously on first render from the persisted tab store so
+  // session-restored tabs don't re-emit on first click.
+  const observedTabsRef = useRef<Map<string, string> | null>(null);
+  if (observedTabsRef.current === null) {
+    const seed = new Map<string, string>();
+    for (const [slug, group] of Object.entries(useTabStore.getState().byWorkspace)) {
+      for (const tab of group.tabs) {
+        seed.set(`${slug}:${tab.id}`, tab.path);
+      }
+    }
+    observedTabsRef.current = seed;
+  }
+  const lastSurfaceRef = useRef<{
+    kind: "login" | "overlay" | "tab" | null;
+    key: string | null;
+    path: string | null;
+  }>({ kind: null, key: null, path: null });
+
+  useEffect(() => {
+    let kind: "login" | "overlay" | "tab";
+    let path: string;
+    let key: string | null = null;
+
+    if (!user) {
+      kind = "login";
+      path = "/login";
+    } else if (overlay) {
+      kind = "overlay";
+      path = overlayPath(overlay);
+    } else if (activeTabPath && activeTabId && activeWorkspaceSlug) {
+      kind = "tab";
+      key = `${activeWorkspaceSlug}:${activeTabId}`;
+      path = activeTabPath;
+    } else {
+      return;
+    }
+
+    const observed = observedTabsRef.current!;
+    const last = lastSurfaceRef.current;
+    const next = { kind, key, path };
+
+    if (kind === "tab" && key !== null) {
+      const knownPath = observed.get(key);
+      const isReactivation =
+        last.key !== key && knownPath !== undefined && knownPath === path;
+      observed.set(key, path);
+      if (isReactivation) {
+        lastSurfaceRef.current = next;
+        return;
+      }
+    }
+
+    const unchanged =
+      last.kind === kind && last.key === key && last.path === path;
+    if (unchanged) return;
+
+    capturePageview(path);
+    lastSurfaceRef.current = next;
+  }, [user, overlay, activeWorkspaceSlug, activeTabId, activeTabPath]);
+
+  return null;
+}
+
+function overlayPath(overlay: WindowOverlay): string {
+  switch (overlay.type) {
+    case "new-workspace":
+      return "/workspaces/new";
+    case "onboarding":
+      return "/onboarding";
+    case "invite":
+      return `/invite/${overlay.invitationId}`;
+    case "invitations":
+      return "/invitations";
+  }
+}

apps/desktop/src/renderer/src/components/parse-daemon-log.test.ts

@@ -0,0 +1,124 @@
+import { describe, it, expect } from "vitest";
+import { parseLogLine } from "./parse-daemon-log";
+
+// All sample lines below are taken verbatim from real daemon output (Go
+// `slog` + `lmittmann/tint` v1.1.3 with NoColor=true). The parser must
+// stay aligned with what tint actually writes — not what we assume.
+
+describe("parseLogLine", () => {
+  it("parses tint's 3-letter INF level", () => {
+    const line =
+      "17:52:35.587 INF task completed component=daemon task=c45266e5 status=completed";
+    const r = parseLogLine(line, 1);
+    expect(r.timestamp).toBe("17:52:35.587");
+    expect(r.level).toBe("INFO");
+    expect(r.message).toBe("task completed");
+    expect(r.fields).toEqual({
+      component: "daemon",
+      task: "c45266e5",
+      status: "completed",
+    });
+  });
+
+  it("parses 3-letter DBG / WRN / ERR levels", () => {
+    expect(parseLogLine("17:53:06.644 DBG agent component=daemon", 1).level).toBe("DEBUG");
+    expect(parseLogLine("07:48:09.391 WRN claim task failed component=daemon", 1).level).toBe("WARN");
+    expect(parseLogLine("12:00:00.000 ERR something bad component=daemon", 1).level).toBe("ERROR");
+  });
+
+  it("still accepts 4-letter level names (defensive against config changes)", () => {
+    const r = parseLogLine("12:00:00.000 INFO regular component=daemon", 1);
+    expect(r.level).toBe("INFO");
+    expect(r.message).toBe("regular");
+  });
+
+  it("tolerates the +N / -N delta tint appends for non-standard slog levels", () => {
+    // tint emits e.g. "INF+1" when slog.Log is called with LevelInfo+1.
+    // We treat the base level as canonical and drop the delta from the UI.
+    const r = parseLogLine("12:00:00.000 INF+1 unusual delta component=daemon", 1);
+    expect(r.level).toBe("INFO");
+    expect(r.message).toBe("unusual delta");
+  });
+
+  it("preserves message text containing colons and special chars", () => {
+    // Real sample: "tool #1: Skill component=daemon task=..."
+    const r = parseLogLine(
+      "17:52:54.578 INF tool #1: Skill component=daemon task=8791b717",
+      1,
+    );
+    expect(r.message).toBe("tool #1: Skill");
+    expect(r.fields).toEqual({ component: "daemon", task: "8791b717" });
+  });
+
+  it("unquotes a double-quoted value containing escaped quotes", () => {
+    // Real sample with escaped quotes inside the agent's emitted text.
+    const line =
+      '17:53:06.644 DBG agent component=daemon task=8791b717 text="The issue is just \\"ping\\" with no description."';
+    const r = parseLogLine(line, 1);
+    expect(r.message).toBe("agent");
+    expect(r.fields.text).toBe('The issue is just "ping" with no description.');
+    expect(r.fields.task).toBe("8791b717");
+  });
+
+  it("handles a quoted value containing a URL with embedded escaped quotes and a colon", () => {
+    // Real sample: error="Post \"http://...\": dial tcp ..."
+    const line =
+      '07:48:09.391 WRN claim task failed component=daemon runtime_id=03f8ff17-276d error="Post \\"http://localhost:8080/api/daemon/runtimes/abc/tasks/claim\\": dial tcp [::1]:8080: connect: connection refused"';
+    const r = parseLogLine(line, 1);
+    expect(r.level).toBe("WARN");
+    expect(r.message).toBe("claim task failed");
+    expect(r.fields.runtime_id).toBe("03f8ff17-276d");
+    expect(r.fields.error).toBe(
+      'Post "http://localhost:8080/api/daemon/runtimes/abc/tasks/claim": dial tcp [::1]:8080: connect: connection refused',
+    );
+  });
+
+  it("handles a quoted value with internal whitespace (e.g. args array)", () => {
+    const line =
+      '17:52:48.757 INF agent command component=daemon exec=claude args="[-p --output-format stream-json --verbose]"';
+    const r = parseLogLine(line, 1);
+    expect(r.message).toBe("agent command");
+    expect(r.fields.exec).toBe("claude");
+    expect(r.fields.args).toBe("[-p --output-format stream-json --verbose]");
+  });
+
+  it("handles message words ending with characters before the field block", () => {
+    // 'execenv:' is part of the message — the colon shouldn't confuse parsing.
+    const r = parseLogLine(
+      "17:52:48.757 INF execenv: prepared env component=daemon repos_available=0",
+      1,
+    );
+    expect(r.message).toBe("execenv: prepared env");
+    expect(r.fields).toEqual({ component: "daemon", repos_available: "0" });
+  });
+
+  it("falls back to raw rendering for non-matching lines (panic stack frame)", () => {
+    const r = parseLogLine("\tat github.com/multica/foo (line 42)", 1);
+    expect(r.timestamp).toBeNull();
+    expect(r.level).toBeNull();
+    expect(r.message).toBe("\tat github.com/multica/foo (line 42)");
+    expect(r.fields).toEqual({});
+    expect(r.raw).toBe("\tat github.com/multica/foo (line 42)");
+  });
+
+  it("falls back to raw rendering for unrecognised level tokens", () => {
+    // If tint ever emits something we don't know, never crash; show raw.
+    const r = parseLogLine("12:00:00.000 TRACE something exotic", 1);
+    expect(r.timestamp).toBeNull();
+    expect(r.level).toBeNull();
+    expect(r.raw).toBe("12:00:00.000 TRACE something exotic");
+  });
+
+  it("attaches an id to every parsed line for stable React keys", () => {
+    const a = parseLogLine("17:52:35.587 INF first component=daemon", 7);
+    const b = parseLogLine("17:52:35.588 INF second component=daemon", 8);
+    expect(a.id).toBe(7);
+    expect(b.id).toBe(8);
+  });
+
+  it("returns empty fields object when there are no key=value pairs", () => {
+    const r = parseLogLine("17:52:35.587 INF a bare message with no fields", 1);
+    expect(r.message).toBe("a bare message with no fields");
+    expect(r.fields).toEqual({});
+  });
+});

apps/desktop/src/renderer/src/components/parse-daemon-log.ts

@@ -0,0 +1,96 @@
+// Pure parser for daemon log lines. The daemon writes via Go's slog with
+// the `tint` handler in NoColor mode (the file isn't a TTY), so each line
+// has a stable shape:
+//
+//   HH:MM:SS.mmm  LEVEL  message text  key=value key2="quoted value"
+//
+// We split it into structured pieces so the UI can render timestamp,
+// level, message and structured fields in separate columns and let users
+// filter / search across them. Anything that doesn't match (panic stack
+// traces, third-party prints, partial writes during log rotation) falls
+// back to a raw view — we never drop input.
+
+export type LogLevel = "DEBUG" | "INFO" | "WARN" | "ERROR";
+
+export interface ParsedLogLine {
+  /** Monotonic id assigned at receive time; stable across re-renders. */
+  id: number;
+  /** "HH:MM:SS.mmm" or null when the line didn't match the standard shape. */
+  timestamp: string | null;
+  level: LogLevel | null;
+  /** Human-readable message body, with structured fields stripped off. */
+  message: string;
+  /** key/value pairs trailing the message. Empty if there were none. */
+  fields: Record<string, string>;
+  /** The original line, kept for fallback rendering and copy-to-clipboard. */
+  raw: string;
+}
+
+// `tint` v1.x emits the 3-letter short form (DBG / INF / WRN / ERR) and,
+// for non-standard slog levels, appends a signed delta (e.g. "INF+1",
+// "DBG-2"). We accept both the short and 4-letter long forms (defensive
+// against future config changes) and normalize them to a canonical
+// 4-letter LogLevel. The optional `[+-]\d+` suffix is captured into the
+// regex and discarded — surfacing `INF+1` to the UI doesn't help users
+// and complicates the level filter chips.
+const HEADER_RE =
+  /^(\d{2}:\d{2}:\d{2}\.\d{3})\s+(DEBUG|DBG|INFO|INF|WARN|WRN|ERROR|ERR)(?:[+-]\d+)?\s+(.+)$/;
+
+const LEVEL_NORMALIZE: Record<string, LogLevel> = {
+  DEBUG: "DEBUG",
+  DBG: "DEBUG",
+  INFO: "INFO",
+  INF: "INFO",
+  WARN: "WARN",
+  WRN: "WARN",
+  ERROR: "ERROR",
+  ERR: "ERROR",
+};
+// Anchored to the END of the remaining string so we peel one field at a
+// time from the right. `value` is either a double-quoted string (which may
+// contain escaped chars) or any non-whitespace run.
+const TRAILING_FIELD_RE = /\s+([a-zA-Z_][a-zA-Z0-9_.]*)=("(?:[^"\\]|\\.)*"|\S+)$/;
+
+function unquote(value: string): string {
+  if (value.length >= 2 && value.startsWith('"') && value.endsWith('"')) {
+    return value.slice(1, -1).replace(/\\"/g, '"').replace(/\\\\/g, "\\");
+  }
+  return value;
+}
+
+function extractTrailingFields(rest: string): {
+  message: string;
+  fields: Record<string, string>;
+} {
+  const fields: Record<string, string> = {};
+  let work = rest;
+  while (true) {
+    const match = work.match(TRAILING_FIELD_RE);
+    if (!match || match.index === undefined) break;
+    fields[match[1]!] = unquote(match[2]!);
+    work = work.slice(0, match.index);
+  }
+  return { message: work.trim(), fields };
+}
+
+export function parseLogLine(raw: string, id: number): ParsedLogLine {
+  const match = raw.match(HEADER_RE);
+  if (!match) {
+    return { id, timestamp: null, level: null, message: raw, fields: {}, raw };
+  }
+  const [, timestamp, level, rest] = match;
+  const normalized = LEVEL_NORMALIZE[level!];
+  if (!normalized) {
+    // Unknown level token — keep raw shape so we don't mis-categorize.
+    return { id, timestamp: null, level: null, message: raw, fields: {}, raw };
+  }
+  const { message, fields } = extractTrailingFields(rest!);
+  return {
+    id,
+    timestamp: timestamp!,
+    level: normalized,
+    message,
+    fields,
+    raw,
+  };
+}

apps/desktop/src/renderer/src/components/tab-bar.tsx

@@ -0,0 +1,189 @@
+import {
+  Inbox,
+  CircleUser,
+  ListTodo,
+  Bot,
+  Monitor,
+  BookOpenText,
+  Settings,
+  X,
+  Plus,
+  type LucideIcon,
+} from "lucide-react";
+import {
+  DndContext,
+  PointerSensor,
+  useSensor,
+  useSensors,
+  closestCenter,
+  type DragEndEvent,
+} from "@dnd-kit/core";
+import {
+  SortableContext,
+  horizontalListSortingStrategy,
+  useSortable,
+} from "@dnd-kit/sortable";
+import {
+  restrictToHorizontalAxis,
+  restrictToParentElement,
+} from "@dnd-kit/modifiers";
+import { CSS } from "@dnd-kit/utilities";
+import { cn } from "@multica/ui/lib/utils";
+import { useTabStore, useActiveGroup, resolveRouteIcon, type Tab } from "@/stores/tab-store";
+import { paths } from "@multica/core/paths";
+
+const TAB_ICONS: Record<string, LucideIcon> = {
+  Inbox,
+  CircleUser,
+  ListTodo,
+  Bot,
+  Monitor,
+  BookOpenText,
+  Settings,
+};
+
+function SortableTabItem({ tab, isActive, isOnly }: { tab: Tab; isActive: boolean; isOnly: boolean }) {
+  const setActiveTab = useTabStore((s) => s.setActiveTab);
+  const closeTab = useTabStore((s) => s.closeTab);
+
+  const {
+    attributes,
+    listeners,
+    setNodeRef,
+    transform,
+    transition,
+    isDragging,
+  } = useSortable({ id: tab.id });
+
+  const Icon = TAB_ICONS[tab.icon];
+
+  const style = {
+    transform: CSS.Transform.toString(transform),
+    transition,
+    WebkitAppRegion: "no-drag",
+    zIndex: isDragging ? 10 : undefined,
+  } as React.CSSProperties;
+
+  const handleClick = () => {
+    if (isActive) return;
+    setActiveTab(tab.id);
+  };
+
+  const handleClose = (e: React.MouseEvent) => {
+    e.stopPropagation();
+    closeTab(tab.id);
+  };
+
+  const stopDragOnClose = (e: React.PointerEvent) => {
+    e.stopPropagation();
+  };
+
+  return (
+    <button
+      ref={setNodeRef}
+      style={style}
+      {...attributes}
+      {...listeners}
+      onClick={handleClick}
+      className={cn(
+        "group flex h-7 w-40 items-center gap-1.5 rounded-md px-2 text-xs transition-colors",
+        "select-none cursor-default",
+        isActive
+          ? "bg-sidebar-accent font-medium text-sidebar-accent-foreground"
+          : "bg-sidebar-accent/50 text-muted-foreground hover:bg-sidebar-accent hover:text-sidebar-accent-foreground",
+        isDragging && "opacity-60",
+      )}
+    >
+      {Icon && <Icon className="size-3.5 shrink-0" />}
+      <span
+        className="min-w-0 flex-1 overflow-hidden whitespace-nowrap text-left"
+        style={{
+          maskImage: "linear-gradient(to right, black calc(100% - 12px), transparent)",
+          WebkitMaskImage: "linear-gradient(to right, black calc(100% - 12px), transparent)",
+        }}
+      >
+        {tab.title}
+      </span>
+      {!isOnly && (
+        <span
+          onClick={handleClose}
+          onPointerDown={stopDragOnClose}
+          className="hidden size-3.5 shrink-0 items-center justify-center rounded-sm text-muted-foreground transition-colors group-hover:flex hover:bg-muted-foreground/20 hover:text-foreground"
+        >
+          <X className="size-2.5" />
+        </span>
+      )}
+    </button>
+  );
+}
+
+function NewTabButton() {
+  const addTab = useTabStore((s) => s.addTab);
+  const setActiveTab = useTabStore((s) => s.setActiveTab);
+
+  const handleClick = () => {
+    // New tab opens in the currently active workspace — tabs are scoped
+    // per workspace, so there is no cross-workspace ambiguity to resolve.
+    const activeSlug = useTabStore.getState().activeWorkspaceSlug;
+    if (!activeSlug) return;
+    const path = paths.workspace(activeSlug).issues();
+    const tabId = addTab(path, "Issues", resolveRouteIcon(path));
+    if (tabId) setActiveTab(tabId);
+  };
+
+  return (
+    <button
+      onClick={handleClick}
+      style={{ WebkitAppRegion: "no-drag" } as React.CSSProperties}
+      className="flex size-7 shrink-0 items-center justify-center rounded-md text-muted-foreground/70 transition-colors hover:bg-muted/50 hover:text-muted-foreground"
+    >
+      <Plus className="size-3.5" />
+    </button>
+  );
+}
+
+export function TabBar() {
+  const group = useActiveGroup();
+  const moveTab = useTabStore((s) => s.moveTab);
+
+  const sensors = useSensors(
+    useSensor(PointerSensor, {
+      activationConstraint: { distance: 5 },
+    }),
+  );
+
+  const tabs = group?.tabs ?? [];
+  const activeTabId = group?.activeTabId ?? "";
+  const tabIds = tabs.map((t) => t.id);
+
+  const handleDragEnd = (event: DragEndEvent) => {
+    const { active, over } = event;
+    if (!over || active.id === over.id) return;
+    const from = tabs.findIndex((t) => t.id === active.id);
+    const to = tabs.findIndex((t) => t.id === over.id);
+    if (from !== -1 && to !== -1) moveTab(from, to);
+  };
+
+  return (
+    <div className="flex h-full items-center gap-0.5 px-2 justify-start">
+      <DndContext
+        sensors={sensors}
+        collisionDetection={closestCenter}
+        modifiers={[restrictToHorizontalAxis, restrictToParentElement]}
+        onDragEnd={handleDragEnd}
+      >
+        <SortableContext items={tabIds} strategy={horizontalListSortingStrategy}>
+          {tabs.map((tab) => (
+            <SortableTabItem
+              key={tab.id}
+              tab={tab}
+              isActive={tab.id === activeTabId}
+              isOnly={tabs.length === 1}
+            />
+          ))}
+        </SortableContext>
+      </DndContext>
+      {group && <NewTabButton />}
+    </div>
+  );
+}

apps/desktop/src/renderer/src/components/tab-content.tsx

@@ -0,0 +1,55 @@
+import { Activity, useEffect } from "react";
+import { RouterProvider } from "react-router-dom";
+import { useActiveGroup } from "@/stores/tab-store";
+import { TabNavigationProvider } from "@/platform/navigation";
+import { useTabRouterSync } from "@/hooks/use-tab-router-sync";
+import type { Tab } from "@/stores/tab-store";
+
+/**
+ * Inner wrapper rendered inside each tab's RouterProvider. The router
+ * reference is stable for a tab's lifetime, so passing it in directly
+ * (instead of re-deriving from the store) avoids needless re-renders.
+ */
+function TabRouterInner({ tab }: { tab: Tab }) {
+  useTabRouterSync(tab.id, tab.router);
+  return null;
+}
+
+/**
+ * Renders the active workspace's tabs using Activity for state preservation.
+ * Only the active tab is visible; hidden tabs keep their DOM and React state.
+ *
+ * When switching workspaces, the previous workspace's tabs unmount entirely
+ * and the new workspace's tabs mount fresh — cross-workspace state
+ * preservation is an explicit non-goal (keeping all workspaces' tabs warm
+ * simultaneously would bloat memory and make workspace switching feel
+ * anything but "switching").
+ */
+export function TabContent() {
+  const group = useActiveGroup();
+
+  // Sync document.title when switching tabs within the active workspace.
+  useEffect(() => {
+    if (!group) return;
+    const tab = group.tabs.find((t) => t.id === group.activeTabId);
+    if (tab) document.title = tab.title;
+  }, [group?.activeTabId, group?.tabs]);
+
+  if (!group) return null;
+
+  return (
+    <>
+      {group.tabs.map((tab) => (
+        <Activity
+          key={tab.id}
+          mode={tab.id === group.activeTabId ? "visible" : "hidden"}
+        >
+          <TabNavigationProvider router={tab.router}>
+            <RouterProvider router={tab.router} />
+            <TabRouterInner tab={tab} />
+          </TabNavigationProvider>
+        </Activity>
+      ))}
+    </>
+  );
+}

apps/desktop/src/renderer/src/components/update-notification.tsx

@@ -0,0 +1,137 @@
+import { useCallback, useEffect, useState } from "react";
+import { ArrowDownToLine, RefreshCw, X } from "lucide-react";
+
+type UpdateState =
+  | { status: "idle" }
+  | { status: "available"; version: string }
+  | { status: "downloading"; percent: number }
+  | { status: "ready" };
+
+export function UpdateNotification() {
+  const [state, setState] = useState<UpdateState>({ status: "idle" });
+  const [dismissed, setDismissed] = useState(false);
+
+  useEffect(() => {
+    const cleanups: (() => void)[] = [];
+
+    cleanups.push(
+      window.updater.onUpdateAvailable((info) => {
+        setState({ status: "available", version: info.version });
+        setDismissed(false);
+      }),
+    );
+
+    cleanups.push(
+      window.updater.onDownloadProgress((progress) => {
+        setState({ status: "downloading", percent: progress.percent });
+      }),
+    );
+
+    cleanups.push(
+      window.updater.onUpdateDownloaded(() => {
+        setState({ status: "ready" });
+      }),
+    );
+
+    return () => cleanups.forEach((fn) => fn());
+  }, []);
+
+  const handleDownload = useCallback(() => {
+    // Prevent double-click: immediately transition to downloading state
+    if (state.status !== "available") return;
+    setState({ status: "downloading", percent: 0 });
+    window.updater.downloadUpdate();
+  }, [state.status]);
+
+  const handleInstall = useCallback(() => {
+    window.updater.installUpdate();
+  }, []);
+
+  // Only allow dismiss when update is available (not during download or ready)
+  if (state.status === "idle") return null;
+  if (dismissed && state.status === "available") return null;
+
+  return (
+    <div className="fixed bottom-4 right-4 z-50 w-80 rounded-lg border border-border bg-background p-4 shadow-lg animate-in slide-in-from-bottom-2 fade-in duration-300">
+      <button
+        onClick={() => setDismissed(true)}
+        className="absolute top-2 right-2 rounded-md p-1 text-muted-foreground hover:text-foreground transition-colors"
+      >
+        <X className="size-3.5" />
+      </button>
+
+      {state.status === "available" && (
+        <div className="flex items-start gap-3">
+          <div className="mt-0.5 rounded-md bg-primary/10 p-1.5">
+            <ArrowDownToLine className="size-4 text-primary" />
+          </div>
+          <div className="flex-1 min-w-0">
+            <p className="text-sm font-medium">New version available</p>
+            <p className="text-xs text-muted-foreground mt-0.5">
+              v{state.version} is ready to download
+            </p>
+            <button
+              onClick={handleDownload}
+              className="mt-2 inline-flex items-center rounded-md bg-primary px-3 py-1.5 text-xs font-medium text-primary-foreground hover:bg-primary/90 transition-colors"
+            >
+              Download update
+            </button>
+          </div>
+        </div>
+      )}
+
+      {state.status === "downloading" && (
+        <div className="flex items-start gap-3">
+          <div className="mt-0.5 rounded-md bg-primary/10 p-1.5">
+            <ArrowDownToLine className="size-4 text-primary animate-pulse" />
+          </div>
+          <div className="flex-1 min-w-0">
+            <p className="text-sm font-medium">Downloading update...</p>
+            <div className="mt-2 h-1.5 w-full rounded-full bg-muted overflow-hidden">
+              <div
+                className="h-full rounded-full bg-primary transition-all duration-300"
+                style={{ width: `${Math.round(state.percent)}%` }}
+              />
+            </div>
+            <p className="text-xs text-muted-foreground mt-1">
+              {Math.round(state.percent)}%
+            </p>
+          </div>
+        </div>
+      )}
+
+      {state.status === "ready" && (
+        <div className="flex items-start gap-3">
+          <div className="mt-0.5 rounded-md bg-success/10 p-1.5">
+            <RefreshCw className="size-4 text-success" />
+          </div>
+          <div className="flex-1 min-w-0">
+            <p className="text-sm font-medium">Update ready</p>
+            <p className="text-xs text-muted-foreground mt-0.5">
+              Restart to apply the update
+            </p>
+            <div className="mt-2 flex items-center gap-1.5">
+              {/* Secondary "See changes" — gives the user a reason to
+                  restart by surfacing what they're about to get. Opens
+                  in the default browser via the shared openExternal
+                  bridge so the URL hits the same allow-list as every
+                  other outbound link. */}
+              <button
+                onClick={() => window.desktopAPI.openExternal("https://multica.ai/changelog")}
+                className="inline-flex items-center rounded-md border border-border bg-background px-3 py-1.5 text-xs font-medium text-foreground hover:bg-accent transition-colors"
+              >
+                See changes
+              </button>
+              <button
+                onClick={handleInstall}
+                className="inline-flex items-center rounded-md bg-primary px-3 py-1.5 text-xs font-medium text-primary-foreground hover:bg-primary/90 transition-colors"
+              >
+                Restart now
+              </button>
+            </div>
+          </div>
+        </div>
+      )}
+    </div>
+  );
+}

apps/desktop/src/renderer/src/components/updates-settings-tab.tsx

@@ -0,0 +1,96 @@
+import { useCallback, useState } from "react";
+import { AlertCircle, ArrowDownToLine, Check, Loader2 } from "lucide-react";
+import { Button } from "@multica/ui/components/ui/button";
+
+type CheckState =
+  | { status: "idle" }
+  | { status: "checking" }
+  | { status: "up-to-date" }
+  | { status: "available"; latestVersion: string }
+  | { status: "error"; message: string };
+
+export function UpdatesSettingsTab() {
+  const [state, setState] = useState<CheckState>({ status: "idle" });
+  const currentVersion = window.desktopAPI.appInfo.version;
+
+  const handleCheck = useCallback(async () => {
+    setState({ status: "checking" });
+    const result = await window.updater.checkForUpdates();
+    if (!result.ok) {
+      setState({ status: "error", message: result.error });
+      return;
+    }
+    setState(
+      result.available
+        ? { status: "available", latestVersion: result.latestVersion }
+        : { status: "up-to-date" },
+    );
+  }, []);
+
+  return (
+    <div>
+      <h2 className="text-lg font-semibold">Updates</h2>
+      <p className="text-sm text-muted-foreground mt-1">
+        The desktop app checks for new versions automatically once an hour and
+        shortly after launch.
+      </p>
+
+      <div className="mt-6 divide-y">
+        <div className="flex items-center justify-between gap-6 py-4">
+          <div className="min-w-0">
+            <p className="text-sm font-medium">Current version</p>
+            <p className="text-sm text-muted-foreground mt-0.5 font-mono">
+              v{currentVersion}
+            </p>
+          </div>
+        </div>
+
+        <div className="flex items-start justify-between gap-6 py-4">
+          <div className="min-w-0">
+            <p className="text-sm font-medium">Check for updates</p>
+            <p className="text-sm text-muted-foreground mt-0.5">
+              Trigger a check now instead of waiting for the next automatic
+              poll. Available updates appear as a notification in the corner.
+            </p>
+            {state.status === "up-to-date" && (
+              <p className="text-sm text-muted-foreground mt-2 inline-flex items-center gap-1.5">
+                <Check className="size-3.5 text-success" />
+                You&apos;re on the latest version.
+              </p>
+            )}
+            {state.status === "available" && (
+              <p className="text-sm text-muted-foreground mt-2 inline-flex items-center gap-1.5">
+                <ArrowDownToLine className="size-3.5 text-primary" />
+                v{state.latestVersion} is available — see the download prompt
+                in the corner.
+              </p>
+            )}
+            {state.status === "error" && (
+              <p className="text-sm text-destructive mt-2 inline-flex items-center gap-1.5">
+                <AlertCircle className="size-3.5" />
+                {state.message}
+              </p>
+            )}
+          </div>
+          <div className="shrink-0">
+            <Button
+              variant="outline"
+              size="sm"
+              onClick={handleCheck}
+              disabled={state.status === "checking"}
+            >
+              {state.status === "checking" ? (
+                <>
+                  <Loader2 className="size-3.5 animate-spin" />
+                  Checking…
+                </>
+              ) : (
+                "Check now"
+              )}
+            </Button>
+          </div>
+        </div>
+      </div>
+    </div>
+  );
+}

apps/desktop/src/renderer/src/components/window-overlay.tsx

@@ -0,0 +1,81 @@
+import { useQuery } from "@tanstack/react-query";
+import { NewWorkspacePage } from "@multica/views/workspace/new-workspace-page";
+import { InvitePage } from "@multica/views/invite";
+import { InvitationsPage } from "@multica/views/invitations";
+import { OnboardingFlow } from "@multica/views/onboarding";
+import { useNavigation } from "@multica/views/navigation";
+import { paths } from "@multica/core/paths";
+import { workspaceListOptions } from "@multica/core/workspace/queries";
+import { useWindowOverlayStore } from "@/stores/window-overlay-store";
+
+/**
+ * Window-level transition overlay: renders above the tab system when the
+ * user is in a pre-workspace flow (onboarding, create workspace, accept
+ * invite).
+ *
+ * This component is intentionally thin — just a fixed positioning shell
+ * that covers the tab system. It does NOT hide traffic lights or provide
+ * a drag strip: each contained view (OnboardingFlow, NewWorkspacePage,
+ * InvitePage) renders its own `<DragStrip />` as a flex-child at top so
+ * native macOS traffic lights stay visible and the page content can fill
+ * the window edge-to-edge. This matches the Linear/Notion/Arc pattern for
+ * pre-dashboard flows and keeps platform chrome consistent across every
+ * "not-in-dashboard" surface.
+ *
+ * All UX affordances (Back button, Log out button, welcome copy, invite
+ * card) live inside the shared view components under `packages/views/`,
+ * so web and desktop render identical content.
+ */
+export function WindowOverlay() {
+  const overlay = useWindowOverlayStore((s) => s.overlay);
+  if (!overlay) return null;
+  return <WindowOverlayInner />;
+}
+
+function WindowOverlayInner() {
+  const overlay = useWindowOverlayStore((s) => s.overlay);
+  const close = useWindowOverlayStore((s) => s.close);
+  const { push } = useNavigation();
+  const { data: wsList = [] } = useQuery(workspaceListOptions());
+
+  if (!overlay) return null;
+
+  // Back is only meaningful when there's somewhere to go — i.e. the user
+  // has at least one workspace. Zero-workspace users can only Log out or
+  // complete the flow.
+  const onBack = wsList.length > 0 ? close : undefined;
+
+  return (
+    <div className="fixed inset-0 z-50 flex flex-col overflow-auto bg-background">
+      {overlay.type === "new-workspace" && (
+        <NewWorkspacePage
+          onSuccess={(ws) => push(paths.workspace(ws.slug).issues())}
+          onBack={onBack}
+        />
+      )}
+      {overlay.type === "invite" && (
+        <InvitePage
+          invitationId={overlay.invitationId}
+          onBack={onBack}
+        />
+      )}
+      {overlay.type === "invitations" && <InvitationsPage />}
+      {overlay.type === "onboarding" && (
+        <OnboardingFlow
+          onComplete={(ws) => {
+            close();
+            // Post-onboarding landing is always the workspace issues
+            // list. The welcome-issue flow moved into a dialog that
+            // renders on that page (StarterContentPrompt), so the
+            // flow doesn't need to thread a target issue id back here.
+            if (ws) {
+              push(paths.workspace(ws.slug).issues());
+            } else {
+              push(paths.root());
+            }
+          }}
+        />
+      )}
+    </div>
+  );
+}

apps/desktop/src/renderer/src/components/workspace-route-layout.tsx

@@ -0,0 +1,90 @@
+import { useEffect } from "react";
+import { Outlet, useNavigate, useParams } from "react-router-dom";
+import { useQuery } from "@tanstack/react-query";
+import { WorkspaceSlugProvider, paths } from "@multica/core/paths";
+import {
+  workspaceBySlugOptions,
+  workspaceListOptions,
+} from "@multica/core/workspace";
+import { setCurrentWorkspace } from "@multica/core/platform";
+import { useAuthStore } from "@multica/core/auth";
+import { useWorkspaceSeen } from "@multica/views/workspace/use-workspace-seen";
+import { WorkspacePresencePrefetch } from "@multica/views/layout";
+import { useTabStore } from "@/stores/tab-store";
+
+/**
+ * Desktop equivalent of apps/web/app/[workspaceSlug]/layout.tsx.
+ *
+ * Resolves the URL slug → workspace UUID via the React Query list cache
+ * (seeded by AuthInitializer). Children do not render until the workspace
+ * is fully resolved — useWorkspaceId() inside child pages is therefore
+ * guaranteed non-null when called. Two industry-standard identities are
+ * kept distinct: slug (URL / browser) and UUID (API / cache keys).
+ *
+ * Unlike web, desktop never renders a "workspace not available" page: the
+ * app has no URL bar and no clickable links from outside the session, so
+ * landing on an inaccessible slug can only mean stale state (a persisted
+ * tab group for a workspace the current user no longer has access to, or
+ * active eviction). Both cases resolve by dropping the stale tab group
+ * from the tab store — the TabBar then renders a different workspace or
+ * the WindowOverlay takes over (zero valid workspaces).
+ */
+export function WorkspaceRouteLayout() {
+  const { workspaceSlug } = useParams<{ workspaceSlug: string }>();
+  const navigate = useNavigate();
+  const user = useAuthStore((s) => s.user);
+  const isAuthLoading = useAuthStore((s) => s.isLoading);
+
+  // Workspace routes require auth. If user is unauthenticated, bounce to /login.
+  useEffect(() => {
+    if (!isAuthLoading && !user) navigate(paths.login(), { replace: true });
+  }, [isAuthLoading, user, navigate]);
+
+  const { data: workspace, isFetched: listFetched } = useQuery({
+    ...workspaceBySlugOptions(workspaceSlug ?? ""),
+    enabled: !!user && !!workspaceSlug,
+  });
+
+  const { data: wsList } = useQuery({
+    ...workspaceListOptions(),
+    enabled: !!user,
+  });
+
+  // Feed the URL slug into the platform singleton so the API client's
+  // X-Workspace-Slug header and persist namespace follow the active tab.
+  // setCurrentWorkspace self-dedupes on slug equality.
+  if (workspace && workspaceSlug) {
+    setCurrentWorkspace(workspaceSlug, workspace.id);
+  }
+
+  const hasBeenSeen = useWorkspaceSeen(workspaceSlug, !!workspace);
+
+  // Stale-slug auto-heal: when this tab's slug fails to resolve, drop the
+  // whole workspace group from the tab store. Per-workspace tab grouping
+  // means the cleanup is a single validator call — the TabContent will
+  // unmount this tab (and all siblings in the stale group) once the store
+  // updates. We don't navigate this tab's router because the tab's path
+  // is scoped to the stale slug; navigating to "/" would create an
+  // inconsistent "tab in group X with path /" state.
+  useEffect(() => {
+    if (!user) return;
+    if (!listFetched) return;
+    if (workspace) return;
+    if (hasBeenSeen) return; // active eviction in flight — let the other path win
+    if (!wsList) return;
+    const validSlugs = new Set(wsList.map((w) => w.slug));
+    useTabStore.getState().validateWorkspaceSlugs(validSlugs);
+  }, [user, listFetched, workspace, hasBeenSeen, wsList]);
+
+  if (isAuthLoading) return null;
+  if (!workspaceSlug) return null;
+  if (!listFetched) return null;
+  if (!workspace) return null; // auto-heal effect above handles the cleanup
+
+  return (
+    <WorkspaceSlugProvider slug={workspaceSlug}>
+      <WorkspacePresencePrefetch />
+      <Outlet />
+    </WorkspaceSlugProvider>
+  );
+}

apps/desktop/src/renderer/src/env.d.ts

@@ -0,0 +1 @@
+/// <reference types="vite/client" />

apps/desktop/src/renderer/src/globals.css

@@ -0,0 +1,42 @@
+@import "tailwindcss";
+@import "tw-animate-css";
+@import "shadcn/tailwind.css";
+@import "@multica/ui/styles/tokens.css";
+@import "@multica/ui/styles/base.css";
+
+@custom-variant dark (&:is(.dark *));
+
+/* Font stack: Inter for Latin UI text + system Chinese fonts for zh content.
+   Web app uses the same stack via next/font/google in apps/web/app/layout.tsx —
+   keep the CJK fallback tail in sync across both files. The Inter primary family
+   differs by design: next/font produces `__Inter_xxx` (with a synthetic size-adjusted
+   fallback face to prevent FOUT layout shift); desktop uses fontsource's "Inter Variable".
+   Both resolve to Inter glyphs, so rendering is identical in practice.
+   Currently covers English + Simplified Chinese. When ja/ko i18n lands, extend
+   the tail with Hiragino Kaku Gothic ProN / Yu Gothic / Apple SD Gothic Neo / Malgun Gothic.
+   Per-character fallback: Latin chars render with Inter, Chinese chars with
+   PingFang SC (macOS) / Microsoft YaHei (Windows) / Noto Sans CJK SC (Linux).
+
+   Mono font has no explicit CJK fallback: CJK chars in code blocks are inherently
+   non-aligned with a mono grid (Chinese is proportional), so listing CJK fonts
+   would falsely signal alignment guarantees. Browser default fallback handles
+   the rare mixed case correctly. */
+:root {
+  --font-sans: "Inter Variable", "Inter", -apple-system, BlinkMacSystemFont,
+               "Segoe UI", "PingFang SC", "Microsoft YaHei", "Noto Sans CJK SC",
+               sans-serif;
+  --font-serif: "Source Serif 4 Variable", "Source Serif 4", "Iowan Old Style",
+                "Apple Garamond", Baskerville, "Times New Roman", serif;
+  --font-mono: "Geist Mono", ui-monospace, SFMono-Regular, Menlo, Consolas,
+               monospace;
+}
+
+@source "../../../../../packages/ui/**/*.tsx";
+@source "../../../../../packages/core/**/*.{ts,tsx}";
+@source "../../../../../packages/views/**/*.{ts,tsx}";
+@source "./**/*.tsx";
+
+/* Desktop-specific: override sidebar container padding for traffic light layout */
+[data-slot="sidebar-container"] {
+  padding: 0 !important;
+}

apps/desktop/src/renderer/src/hooks/use-document-title.ts

@@ -0,0 +1,8 @@
+import { useEffect } from "react";
+
+/** Sets document.title. The tab system observes this automatically. */
+export function useDocumentTitle(title: string) {
+  useEffect(() => {
+    if (title) document.title = title;
+  }, [title]);
+}

apps/desktop/src/renderer/src/hooks/use-tab-history.ts

@@ -0,0 +1,40 @@
+import { useCallback } from "react";
+import type { DataRouter } from "react-router-dom";
+import { useActiveTabRouter, useActiveTabHistory } from "@/stores/tab-store";
+
+/**
+ * Shared hint map so useTabRouterSync can distinguish back vs forward POP.
+ * Set before calling router.navigate(-1 | 1), read in the synchronous subscription.
+ */
+export const popDirectionHints = new Map<DataRouter, "back" | "forward">();
+
+/**
+ * Per-tab back/forward navigation derived from the active workspace's
+ * active tab.
+ *
+ * Subscribed via primitive selectors so this hook only re-renders when
+ * the numeric history state actually changes — path ticks on the active
+ * tab (which don't shift historyIndex) don't churn the back/forward
+ * buttons.
+ */
+export function useTabHistory() {
+  const router = useActiveTabRouter();
+  const { historyIndex, historyLength } = useActiveTabHistory();
+
+  const canGoBack = historyIndex > 0;
+  const canGoForward = historyIndex < historyLength - 1;
+
+  const goBack = useCallback(() => {
+    if (!router || historyIndex <= 0) return;
+    popDirectionHints.set(router, "back");
+    router.navigate(-1);
+  }, [router, historyIndex]);
+
+  const goForward = useCallback(() => {
+    if (!router || historyIndex >= historyLength - 1) return;
+    popDirectionHints.set(router, "forward");
+    router.navigate(1);
+  }, [router, historyIndex, historyLength]);
+
+  return { canGoBack, canGoForward, goBack, goForward };
+}

apps/desktop/src/renderer/src/hooks/use-tab-router-sync.ts

@@ -0,0 +1,49 @@
+import { useEffect, useRef } from "react";
+import type { DataRouter } from "react-router-dom";
+import { useTabStore, resolveRouteIcon } from "@/stores/tab-store";
+import { popDirectionHints } from "./use-tab-history";
+
+/**
+ * Subscribe to a tab's memory router and sync path + history tracking
+ * back into the tab store.
+ *
+ * Called once per tab inside its RouterProvider subtree.
+ */
+export function useTabRouterSync(tabId: string, router: DataRouter) {
+  const indexRef = useRef(0);
+  const lengthRef = useRef(1);
+
+  useEffect(() => {
+    // Sync initial state
+    const initialPath = router.state.location.pathname;
+    const store = useTabStore.getState();
+    store.updateTab(tabId, { path: initialPath, icon: resolveRouteIcon(initialPath) });
+
+    const unsubscribe = router.subscribe((state) => {
+      const { pathname } = state.location;
+      const action = state.historyAction;
+
+      if (action === "PUSH") {
+        indexRef.current += 1;
+        lengthRef.current = indexRef.current + 1;
+      } else if (action === "POP") {
+        // Determine direction from the hint set by goBack/goForward
+        const hint = popDirectionHints.get(router);
+        popDirectionHints.delete(router);
+        if (hint === "forward") {
+          indexRef.current = Math.min(indexRef.current + 1, lengthRef.current - 1);
+        } else {
+          // Default to back
+          indexRef.current = Math.max(0, indexRef.current - 1);
+        }
+      }
+      // REPLACE: index and length stay the same
+
+      const store = useTabStore.getState();
+      store.updateTab(tabId, { path: pathname, icon: resolveRouteIcon(pathname) });
+      store.updateTabHistory(tabId, indexRef.current, lengthRef.current);
+    });
+
+    return unsubscribe;
+  }, [tabId, router]);
+}

apps/desktop/src/renderer/src/hooks/use-tab-sync.ts

@@ -0,0 +1,32 @@
+import { useEffect } from "react";
+import { useTabStore } from "@/stores/tab-store";
+
+/**
+ * Watches document.title via MutationObserver and updates the active tab's
+ * title. Pages set document.title via TitleSync (route handle.title) or
+ * useDocumentTitle(). This observer picks up the change and syncs it to
+ * the tab store.
+ */
+export function useActiveTitleSync() {
+  useEffect(() => {
+    const observer = new MutationObserver(() => {
+      const title = document.title;
+      if (!title) return;
+      const state = useTabStore.getState();
+      if (!state.activeWorkspaceSlug) return;
+      const group = state.byWorkspace[state.activeWorkspaceSlug];
+      if (!group) return;
+      const activeTab = group.tabs.find((t) => t.id === group.activeTabId);
+      if (activeTab && activeTab.title !== title) {
+        state.updateTab(activeTab.id, { title });
+      }
+    });
+
+    const titleEl = document.querySelector("title");
+    if (titleEl) {
+      observer.observe(titleEl, { childList: true, characterData: true, subtree: true });
+    }
+
+    return () => observer.disconnect();
+  }, []);
+}

apps/desktop/src/renderer/src/main.tsx

@@ -0,0 +1,16 @@
+import ReactDOM from "react-dom/client";
+import App from "./App";
+// Inter variable font covers all weights (100-900) in a single file.
+// Geist Mono kept as-is for code blocks; CJK is handled by system font fallback
+// (see globals.css --font-sans chain). Keep font stack in sync with apps/web/app/layout.tsx.
+import "@fontsource-variable/inter";
+// Editorial serif — matches web's next/font Source_Serif_4. Loaded app-wide so
+// onboarding headings and any future editorial surface can use `font-serif`
+// (see tokens.css @theme inline). Variable font = one file covers all weights.
+import "@fontsource-variable/source-serif-4";
+import "@fontsource-variable/source-serif-4/wght-italic.css";
+import "@fontsource/geist-mono/400.css";
+import "@fontsource/geist-mono/700.css";
+import "./globals.css";
+
+ReactDOM.createRoot(document.getElementById("root")!).render(<App />);

apps/desktop/src/renderer/src/pages/agent-detail-page.tsx

@@ -0,0 +1,18 @@
+import { useParams } from "react-router-dom";
+import { useQuery } from "@tanstack/react-query";
+import { AgentDetailPage as SharedAgentDetailPage } from "@multica/views/agents";
+import { useWorkspaceId } from "@multica/core/hooks";
+import { agentListOptions } from "@multica/core/workspace/queries";
+import { useDocumentTitle } from "@/hooks/use-document-title";
+
+export function AgentDetailPage() {
+  const { id } = useParams<{ id: string }>();
+  const wsId = useWorkspaceId();
+  const { data: agents = [] } = useQuery(agentListOptions(wsId));
+  const agent = agents.find((a) => a.id === id) ?? null;
+
+  useDocumentTitle(agent?.name ?? "Agent");
+
+  if (!id) return null;
+  return <SharedAgentDetailPage agentId={id} />;
+}

apps/desktop/src/renderer/src/pages/autopilot-detail-page.tsx

@@ -0,0 +1,17 @@
+import { useParams } from "react-router-dom";
+import { useQuery } from "@tanstack/react-query";
+import { AutopilotDetailPage as AutopilotDetail } from "@multica/views/autopilots/components";
+import { useWorkspaceId } from "@multica/core/hooks";
+import { autopilotDetailOptions } from "@multica/core/autopilots/queries";
+import { useDocumentTitle } from "@/hooks/use-document-title";
+
+export function AutopilotDetailPage() {
+  const { id } = useParams<{ id: string }>();
+  const wsId = useWorkspaceId();
+  const { data } = useQuery(autopilotDetailOptions(wsId, id!));
+
+  useDocumentTitle(data ? `⚡ ${data.autopilot.title}` : "Autopilot");
+
+  if (!id) return null;
+  return <AutopilotDetail autopilotId={id} />;
+}

apps/desktop/src/renderer/src/pages/issue-detail-page.tsx

@@ -0,0 +1,22 @@
+import { useParams } from "react-router-dom";
+import { useQuery } from "@tanstack/react-query";
+import { IssueDetail } from "@multica/views/issues/components";
+import { ErrorBoundary } from "@multica/ui/components/common/error-boundary";
+import { useWorkspaceId } from "@multica/core/hooks";
+import { issueDetailOptions } from "@multica/core/issues/queries";
+import { useDocumentTitle } from "@/hooks/use-document-title";
+
+export function IssueDetailPage() {
+  const { id } = useParams<{ id: string }>();
+  const wsId = useWorkspaceId();
+  const { data: issue } = useQuery(issueDetailOptions(wsId, id!));
+
+  useDocumentTitle(issue ? `${issue.identifier}: ${issue.title}` : "Issue");
+
+  if (!id) return null;
+  return (
+    <ErrorBoundary resetKeys={[id]}>
+      <IssueDetail issueId={id} />
+    </ErrorBoundary>
+  );
+}

apps/desktop/src/renderer/src/pages/login.tsx

@@ -0,0 +1,38 @@
+import { LoginPage } from "@multica/views/auth";
+import { DragStrip } from "@multica/views/platform";
+import { MulticaIcon } from "@multica/ui/components/common/multica-icon";
+
+function requireRuntimeAppUrl(): string {
+  const runtimeConfig = window.desktopAPI.runtimeConfig;
+  if (!runtimeConfig.ok) {
+    throw new Error(
+      "Invariant violated: DesktopLoginPage rendered before App accepted runtime config",
+    );
+  }
+  return runtimeConfig.config.appUrl;
+}
+
+export function DesktopLoginPage() {
+  const webUrl = requireRuntimeAppUrl();
+  const handleGoogleLogin = () => {
+    // Open web login page in the default browser with platform=desktop flag.
+    // The web callback will redirect back via multica:// deep link with the token.
+    window.desktopAPI.openExternal(
+      `${webUrl}/login?platform=desktop`,
+    );
+  };
+
+  return (
+    <div className="flex h-screen flex-col">
+      <DragStrip />
+      <LoginPage
+        logo={<MulticaIcon bordered size="lg" />}
+        onSuccess={() => {
+          // Auth store update triggers AppContent re-render → shows DesktopShell.
+          // Initial workspace navigation happens in routes.tsx via IndexRedirect.
+        }}
+        onGoogleLogin={handleGoogleLogin}
+      />
+    </div>
+  );
+}

apps/desktop/src/renderer/src/pages/project-detail-page.tsx

@@ -0,0 +1,17 @@
+import { useParams } from "react-router-dom";
+import { useQuery } from "@tanstack/react-query";
+import { ProjectDetail } from "@multica/views/projects/components";
+import { useWorkspaceId } from "@multica/core/hooks";
+import { projectDetailOptions } from "@multica/core/projects/queries";
+import { useDocumentTitle } from "@/hooks/use-document-title";
+
+export function ProjectDetailPage() {
+  const { id } = useParams<{ id: string }>();
+  const wsId = useWorkspaceId();
+  const { data: project } = useQuery(projectDetailOptions(wsId, id!));
+
+  useDocumentTitle(project ? `${project.icon || "📁"} ${project.title}` : "Project");
+
+  if (!id) return null;
+  return <ProjectDetail projectId={id} />;
+}

apps/desktop/src/renderer/src/pages/runtime-detail-page.tsx

@@ -0,0 +1,18 @@
+import { useParams } from "react-router-dom";
+import { useQuery } from "@tanstack/react-query";
+import { RuntimeDetailPage as SharedRuntimeDetailPage } from "@multica/views/runtimes";
+import { useWorkspaceId } from "@multica/core/hooks";
+import { runtimeListOptions } from "@multica/core/runtimes/queries";
+import { useDocumentTitle } from "@/hooks/use-document-title";
+
+export function RuntimeDetailPage() {
+  const { id } = useParams<{ id: string }>();
+  const wsId = useWorkspaceId();
+  const { data: runtimes } = useQuery(runtimeListOptions(wsId));
+  const runtime = runtimes?.find((r) => r.id === id);
+
+  useDocumentTitle(runtime?.name ?? "Runtime");
+
+  if (!id) return null;
+  return <SharedRuntimeDetailPage runtimeId={id} />;
+}

apps/desktop/src/renderer/src/pages/skill-detail-page.tsx

@@ -0,0 +1,17 @@
+import { useParams } from "react-router-dom";
+import { useQuery } from "@tanstack/react-query";
+import { SkillDetailPage as SharedSkillDetailPage } from "@multica/views/skills";
+import { useWorkspaceId } from "@multica/core/hooks";
+import { skillDetailOptions } from "@multica/core/workspace/queries";
+import { useDocumentTitle } from "@/hooks/use-document-title";
+
+export function SkillDetailPage() {
+  const { id } = useParams<{ id: string }>();
+  const wsId = useWorkspaceId();
+  const { data: skill } = useQuery(skillDetailOptions(wsId, id ?? ""));
+
+  useDocumentTitle(skill?.name ?? "Skill");
+
+  if (!id) return null;
+  return <SharedSkillDetailPage skillId={id} />;
+}

apps/desktop/src/renderer/src/platform/daemon-ipc-bridge.ts

@@ -0,0 +1,76 @@
+"use client";
+
+import { useEffect } from "react";
+import { useQueryClient } from "@tanstack/react-query";
+import { runtimeKeys } from "@multica/core/runtimes";
+import type { AgentRuntime } from "@multica/core/types";
+
+/**
+ * DesktopAPI exposes a richer DaemonStatus shape than the public AgentRuntime
+ * type — we redeclare the fields we consume here to avoid coupling the bridge
+ * to the desktop preload typings (which live in apps/desktop/src/preload).
+ */
+interface DaemonStatusLike {
+  state: "running" | "stopped" | "starting" | "stopping" | "installing_cli" | "cli_not_found";
+  daemonId?: string;
+}
+
+/**
+ * Merges a local DaemonStatus into an AgentRuntime row. Only the `status`
+ * field is overridden; other fields (name, provider, last_seen_at, etc)
+ * remain server-authoritative. We deliberately ignore intermediate states
+ * (starting / stopping / installing_cli / cli_not_found) so the cache
+ * doesn't flap during boot — if the daemon is in such a state, the runtime
+ * is effectively offline anyway, and the server-side sweeper will mark it
+ * within 75s.
+ */
+function mergeDaemonStatus(rt: AgentRuntime, status: DaemonStatusLike): AgentRuntime {
+  if (status.state === "stopped" || status.state === "stopping") {
+    return { ...rt, status: "offline" };
+  }
+  if (status.state === "running") {
+    return {
+      ...rt,
+      status: "online",
+      last_seen_at: new Date().toISOString(),
+    };
+  }
+  return rt;
+}
+
+/**
+ * Subscribes to local daemon status changes via Electron IPC and writes them
+ * into the runtimes Query cache for the active workspace.
+ *
+ * Why: the server-side runtime sweeper takes up to 75s to flip a runtime to
+ * offline (heartbeat timeout 45s + sweep interval 30s). On the desktop app
+ * we know about local daemon state instantly via IPC, so we use it to
+ * pre-populate the cache and give users a sub-second feedback loop. Web and
+ * "looking at someone else's daemon" still go through the server path.
+ *
+ * Same-daemon-multiple-runtimes: a single daemon can back several runtimes
+ * in the same workspace (one per provider). We map across all matches so
+ * every related runtime row sees the same status flip.
+ */
+export function useDaemonIPCBridge(wsId: string | undefined): void {
+  const qc = useQueryClient();
+
+  useEffect(() => {
+    if (!wsId) return;
+    if (typeof window === "undefined") return;
+    const daemonAPI = (window as unknown as { daemonAPI?: { onStatusChange?: (cb: (s: DaemonStatusLike) => void) => () => void } }).daemonAPI;
+    if (!daemonAPI?.onStatusChange) return;
+
+    const unsubscribe = daemonAPI.onStatusChange((status) => {
+      if (!status.daemonId) return;
+      qc.setQueryData<AgentRuntime[]>(runtimeKeys.list(wsId), (old) => {
+        if (!old) return old;
+        return old.map((rt) =>
+          rt.daemon_id === status.daemonId ? mergeDaemonStatus(rt, status) : rt,
+        );
+      });
+    });
+
+    return unsubscribe;
+  }, [wsId, qc]);
+}

apps/desktop/src/renderer/src/platform/i18n-adapter.ts

@@ -0,0 +1,31 @@
+import type { LocaleAdapter, SupportedLocale } from "@multica/core/i18n";
+
+const STORAGE_KEY = "multica-locale";
+
+// Desktop adapter:
+//   - User choice: localStorage (set by Settings switcher).
+//   - System preference: locale main injected via additionalArguments
+//     (read from preload, exposed on window.desktopAPI.systemLocale).
+//   - Persist: localStorage. The Settings switcher additionally PATCHes
+//     /api/me when logged in so user.language follows the user across devices.
+export function createDesktopLocaleAdapter(systemLocale: string): LocaleAdapter {
+  return {
+    getUserChoice() {
+      try {
+        return window.localStorage.getItem(STORAGE_KEY);
+      } catch {
+        return null;
+      }
+    },
+    getSystemPreferences() {
+      return systemLocale ? [systemLocale] : [];
+    },
+    persist(locale: SupportedLocale) {
+      try {
+        window.localStorage.setItem(STORAGE_KEY, locale);
+      } catch {
+        // Best-effort
+      }
+    },
+  };
+}

apps/desktop/src/renderer/src/platform/navigation.tsx

@@ -0,0 +1,261 @@
+import { useEffect, useMemo, useState } from "react";
+import type { DataRouter } from "react-router-dom";
+import {
+  NavigationProvider,
+  type NavigationAdapter,
+} from "@multica/views/navigation";
+import { useAuthStore } from "@multica/core/auth";
+import { isReservedSlug } from "@multica/core/paths";
+import {
+  useTabStore,
+  resolveRouteIcon,
+  useActiveTabIdentity,
+  useActiveTabRouter,
+  getActiveTab,
+} from "@/stores/tab-store";
+import { useWindowOverlayStore } from "@/stores/window-overlay-store";
+
+function requireRuntimeAppUrl(scope: string): string {
+  const runtimeConfig = window.desktopAPI.runtimeConfig;
+  if (!runtimeConfig.ok) {
+    throw new Error(
+      `Invariant violated: ${scope} rendered before App accepted runtime config`,
+    );
+  }
+  return runtimeConfig.config.appUrl;
+}
+
+/**
+ * Extract the leading workspace slug from a path, or null if the path isn't
+ * workspace-scoped (root, login, any reserved prefix).
+ */
+function extractWorkspaceSlug(path: string): string | null {
+  const first = path.split("/").filter(Boolean)[0] ?? "";
+  if (!first) return null;
+  if (isReservedSlug(first)) return null;
+  return first;
+}
+
+/**
+ * Intercept navigation to "transition" paths — pre-workspace flows that on
+ * desktop are rendered as a window-level overlay instead of a tab route.
+ * Returns `true` if the navigation was handled (caller should NOT proceed).
+ *
+ * Side effect: when opening the new-workspace overlay, the tab router is
+ * ALSO reset to "/". Rationale — the only way a push lands on
+ * /workspaces/new is that the workspace context is gone (fresh install,
+ * delete-last, leave-last). Leaving the tab parked on a workspace-scoped
+ * path would keep those components mounted under the overlay; the next
+ * render after the list cache updates would then throw (useWorkspaceId
+ * etc) because the slug no longer resolves.
+ */
+function tryRouteToOverlay(path: string, router?: DataRouter): boolean {
+  const overlay = useWindowOverlayStore.getState();
+  if (path === "/workspaces/new") {
+    overlay.open({ type: "new-workspace" });
+    if (router && router.state.location.pathname !== "/") {
+      router.navigate("/", { replace: true });
+    }
+    return true;
+  }
+  if (path === "/onboarding") {
+    overlay.open({ type: "onboarding" });
+    if (router && router.state.location.pathname !== "/") {
+      router.navigate("/", { replace: true });
+    }
+    return true;
+  }
+  if (path === "/invitations") {
+    overlay.open({ type: "invitations" });
+    if (router && router.state.location.pathname !== "/") {
+      router.navigate("/", { replace: true });
+    }
+    return true;
+  }
+  if (path.startsWith("/invite/")) {
+    let id = "";
+    try {
+      id = decodeURIComponent(path.slice("/invite/".length));
+    } catch {
+      return true;
+    }
+    if (id) {
+      overlay.open({ type: "invite", invitationId: id });
+      return true;
+    }
+  }
+  // Any other navigation cancels a live overlay.
+  if (overlay.overlay) overlay.close();
+  return false;
+}
+
+/**
+ * Intercept pushes that change workspace. Returns `true` if the navigation
+ * was delegated to the tab store (caller should NOT proceed).
+ *
+ * This is the entry point that makes shared code platform-agnostic:
+ * sidebar dropdown, cmd+k "switch workspace", post-delete redirects,
+ * invite-accept flow — they all call `useNavigation().push(path)` with a
+ * full workspace URL, and on desktop we translate "target slug differs
+ * from active" into "switch the tab-group that's visible in the TabBar".
+ */
+function tryRouteToOtherWorkspace(path: string): boolean {
+  const targetSlug = extractWorkspaceSlug(path);
+  if (!targetSlug) return false;
+  const { activeWorkspaceSlug, switchWorkspace } = useTabStore.getState();
+  if (targetSlug === activeWorkspaceSlug) return false;
+  switchWorkspace(targetSlug, path);
+  return true;
+}
+
+/**
+ * Root-level navigation provider for components outside the per-tab
+ * RouterProviders (sidebar, search dialog, modals, WindowOverlay contents).
+ *
+ * Reads from the active tab's memory router via router.subscribe().
+ * Does NOT use any react-router hooks — it's above all RouterProviders.
+ */
+export function DesktopNavigationProvider({
+  children,
+}: {
+  children: React.ReactNode;
+}) {
+  const appUrl = requireRuntimeAppUrl("DesktopNavigationProvider");
+  // Primitive-only subscriptions so this component doesn't re-render on
+  // unrelated store updates (e.g. an inactive tab's router tick). We
+  // resolve the active router here only to subscribe once per tab switch.
+  const { tabId: activeTabId } = useActiveTabIdentity();
+  const router = useActiveTabRouter();
+  // Mirror the active tab router's full location (pathname + search) so
+  // shell-level consumers of useNavigation() — ChatWindow in particular —
+  // can read URL search params. Must stay in sync with TabNavigationProvider
+  // below; a partial shape here (just pathname) silently broke focus-mode
+  // anchor resolution on `/inbox?issue=…`.
+  const [location, setLocation] = useState<{ pathname: string; search: string }>(
+    () => ({
+      pathname: router?.state.location.pathname ?? "/",
+      search: router?.state.location.search ?? "",
+    }),
+  );
+
+  useEffect(() => {
+    if (!router) {
+      setLocation({ pathname: "/", search: "" });
+      return;
+    }
+    setLocation({
+      pathname: router.state.location.pathname,
+      search: router.state.location.search,
+    });
+    return router.subscribe((state) => {
+      setLocation({
+        pathname: state.location.pathname,
+        search: state.location.search,
+      });
+    });
+  }, [activeTabId, router]);
+
+  const adapter: NavigationAdapter = useMemo(
+    () => ({
+      push: (path: string) => {
+        if (path === "/login") {
+          useAuthStore.getState().logout();
+          return;
+        }
+        const active = currentActiveTab();
+        if (tryRouteToOverlay(path, active?.router)) return;
+        if (tryRouteToOtherWorkspace(path)) return;
+        active?.router.navigate(path);
+      },
+      replace: (path: string) => {
+        const active = currentActiveTab();
+        if (tryRouteToOverlay(path, active?.router)) return;
+        if (tryRouteToOtherWorkspace(path)) return;
+        active?.router.navigate(path, { replace: true });
+      },
+      back: () => {
+        currentActiveTab()?.router.navigate(-1);
+      },
+      pathname: location.pathname,
+      searchParams: new URLSearchParams(location.search),
+      openInNewTab: (path: string, title?: string) => {
+        // Cross-workspace "open in new tab" switches workspace and opens
+        // the path there; same-workspace just adds a tab in the current group.
+        const slug = extractWorkspaceSlug(path);
+        const store = useTabStore.getState();
+        if (slug && slug !== store.activeWorkspaceSlug) {
+          store.switchWorkspace(slug, path);
+          return;
+        }
+        const icon = resolveRouteIcon(path);
+        const tabId = store.openTab(path, title ?? path, icon);
+        if (tabId) store.setActiveTab(tabId);
+      },
+      getShareableUrl: (path: string) => `${appUrl}${path}`,
+    }),
+    [appUrl, location],
+  );
+
+  return <NavigationProvider value={adapter}>{children}</NavigationProvider>;
+}
+
+function currentActiveTab() {
+  return getActiveTab(useTabStore.getState());
+}
+
+/**
+ * Per-tab navigation provider rendered inside each tab's Activity wrapper.
+ * Subscribes to the tab's own router for up-to-date pathname.
+ *
+ * This is what @multica/views page components read via useNavigation().
+ */
+export function TabNavigationProvider({
+  router,
+  children,
+}: {
+  router: DataRouter;
+  children: React.ReactNode;
+}) {
+  const appUrl = requireRuntimeAppUrl("TabNavigationProvider");
+  const [location, setLocation] = useState(router.state.location);
+
+  useEffect(() => {
+    setLocation(router.state.location);
+    return router.subscribe((state) => {
+      setLocation(state.location);
+    });
+  }, [router]);
+
+  const adapter: NavigationAdapter = useMemo(
+    () => ({
+      push: (path: string) => {
+        if (tryRouteToOverlay(path, router)) return;
+        if (tryRouteToOtherWorkspace(path)) return;
+        router.navigate(path);
+      },
+      replace: (path: string) => {
+        if (tryRouteToOverlay(path, router)) return;
+        if (tryRouteToOtherWorkspace(path)) return;
+        router.navigate(path, { replace: true });
+      },
+      back: () => router.navigate(-1),
+      pathname: location.pathname,
+      searchParams: new URLSearchParams(location.search),
+      openInNewTab: (path: string, title?: string) => {
+        const slug = extractWorkspaceSlug(path);
+        const store = useTabStore.getState();
+        if (slug && slug !== store.activeWorkspaceSlug) {
+          store.switchWorkspace(slug, path);
+          return;
+        }
+        const icon = resolveRouteIcon(path);
+        const tabId = store.openTab(path, title ?? path, icon);
+        if (tabId) store.setActiveTab(tabId);
+      },
+      getShareableUrl: (path: string) => `${appUrl}${path}`,
+    }),
+    [appUrl, router, location],
+  );
+
+  return <NavigationProvider value={adapter}>{children}</NavigationProvider>;
+}

apps/desktop/src/renderer/src/routes.tsx

@@ -0,0 +1,182 @@
+import { useEffect } from "react";
+import {
+  createMemoryRouter,
+  Navigate,
+  Outlet,
+  useMatches,
+} from "react-router-dom";
+import type { RouteObject } from "react-router-dom";
+import { IssueDetailPage } from "./pages/issue-detail-page";
+import { ProjectDetailPage } from "./pages/project-detail-page";
+import { AutopilotDetailPage } from "./pages/autopilot-detail-page";
+import { SkillDetailPage } from "./pages/skill-detail-page";
+import { AgentDetailPage } from "./pages/agent-detail-page";
+import { RuntimeDetailPage } from "./pages/runtime-detail-page";
+import { IssuesPage } from "@multica/views/issues/components";
+import { ProjectsPage } from "@multica/views/projects/components";
+import { AutopilotsPage } from "@multica/views/autopilots/components";
+import { MyIssuesPage } from "@multica/views/my-issues";
+import { SkillsPage } from "@multica/views/skills";
+import { DesktopRuntimesPage } from "./components/desktop-runtimes-page";
+import { AgentsPage } from "@multica/views/agents";
+import { InboxPage } from "@multica/views/inbox";
+import { SettingsPage } from "@multica/views/settings";
+import { ErrorBoundary } from "@multica/ui/components/common/error-boundary";
+import { Download, Server } from "lucide-react";
+import { DaemonSettingsTab } from "./components/daemon-settings-tab";
+import { UpdatesSettingsTab } from "./components/updates-settings-tab";
+import { WorkspaceRouteLayout } from "./components/workspace-route-layout";
+
+/**
+ * Sets document.title from the deepest matched route's handle.title.
+ * The tab system observes document.title via MutationObserver.
+ * Pages with dynamic titles (e.g. issue detail) override by setting
+ * document.title directly via useDocumentTitle().
+ */
+function TitleSync() {
+  const matches = useMatches();
+  const title = [...matches]
+    .reverse()
+    .find((m) => (m.handle as { title?: string })?.title)
+    ?.handle as { title?: string } | undefined;
+
+  useEffect(() => {
+    if (title?.title) document.title = title.title;
+  }, [title?.title]);
+
+  return null;
+}
+
+/** Wrapper that renders route children + TitleSync */
+function PageShell() {
+  return (
+    <>
+      <TitleSync />
+      <Outlet />
+    </>
+  );
+}
+
+/**
+ * Route definitions shared by all tabs.
+ *
+ * Every tab path is workspace-scoped: `/{slug}/{route}/...`. Pre-workspace
+ * flows (create workspace, accept invite) are NOT routes — they render as a
+ * window-level overlay via `WindowOverlay`, dispatched by the navigation
+ * adapter's transition-path interception. The `activeWorkspaceSlug` in the
+ * tab store decides which workspace's tabs are visible in the TabBar;
+ * workspace-less state (zero-workspace user) shows the overlay instead.
+ *
+ * The root index route stays as a harmless safety net. With per-workspace
+ * tabs, nothing should construct a tab at `/` — but if one ever slips
+ * through (malformed persisted state that dodges the migration, direct
+ * router.navigate from unforeseen code), the index falls back to null
+ * rather than 404; App.tsx's bootstrap repoints activeWorkspaceSlug on the
+ * next render pass.
+ */
+export const appRoutes: RouteObject[] = [
+  {
+    element: <PageShell />,
+    children: [
+      { index: true, element: null },
+      {
+        path: ":workspaceSlug",
+        element: <WorkspaceRouteLayout />,
+        children: [
+          { index: true, element: <Navigate to="issues" replace /> },
+          {
+            path: "issues",
+            element: (
+              <ErrorBoundary>
+                <IssuesPage />
+              </ErrorBoundary>
+            ),
+            handle: { title: "Issues" },
+          },
+          {
+            path: "issues/:id",
+            element: <IssueDetailPage />,
+            handle: { title: "Issue" },
+          },
+          {
+            path: "projects",
+            element: <ProjectsPage />,
+            handle: { title: "Projects" },
+          },
+          {
+            path: "projects/:id",
+            element: <ProjectDetailPage />,
+            handle: { title: "Project" },
+          },
+          {
+            path: "autopilots",
+            element: <AutopilotsPage />,
+            handle: { title: "Autopilot" },
+          },
+          {
+            path: "autopilots/:id",
+            element: <AutopilotDetailPage />,
+            handle: { title: "Autopilot" },
+          },
+          {
+            path: "my-issues",
+            element: <MyIssuesPage />,
+            handle: { title: "My Issues" },
+          },
+          {
+            path: "runtimes",
+            element: <DesktopRuntimesPage />,
+            handle: { title: "Runtimes" },
+          },
+          {
+            path: "runtimes/:id",
+            element: <RuntimeDetailPage />,
+            handle: { title: "Runtime" },
+          },
+          { path: "skills", element: <SkillsPage />, handle: { title: "Skills" } },
+          {
+            path: "skills/:id",
+            element: <SkillDetailPage />,
+            handle: { title: "Skill" },
+          },
+          { path: "agents", element: <AgentsPage />, handle: { title: "Agents" } },
+          {
+            path: "agents/:id",
+            element: <AgentDetailPage />,
+            handle: { title: "Agent" },
+          },
+          { path: "inbox", element: <InboxPage />, handle: { title: "Inbox" } },
+          {
+            path: "settings",
+            element: (
+              <SettingsPage
+                extraAccountTabs={[
+                  {
+                    value: "daemon",
+                    label: "Daemon",
+                    icon: Server,
+                    content: <DaemonSettingsTab />,
+                  },
+                  {
+                    value: "updates",
+                    label: "Updates",
+                    icon: Download,
+                    content: <UpdatesSettingsTab />,
+                  },
+                ]}
+              />
+            ),
+            handle: { title: "Settings" },
+          },
+        ],
+      },
+    ],
+  },
+];
+
+/** Create an independent memory router for a tab. */
+export function createTabRouter(initialPath: string) {
+  return createMemoryRouter(appRoutes, {
+    initialEntries: [initialPath],
+  });
+}

apps/desktop/src/renderer/src/stores/tab-store.test.ts

@@ -0,0 +1,224 @@
+import { describe, expect, it, vi, beforeEach } from "vitest";
+
+// createTabRouter transitively pulls in route modules that expect a browser
+// router context. For pure store tests we stub it to a minimal disposable.
+const createTabRouterMock = vi.hoisted(() =>
+  vi.fn(() => ({
+    dispose: vi.fn(),
+    state: { location: { pathname: "/" } },
+    navigate: vi.fn(),
+    subscribe: vi.fn(() => () => {}),
+  })),
+);
+vi.mock("../routes", () => ({
+  createTabRouter: createTabRouterMock,
+}));
+
+import {
+  sanitizeTabPath,
+  migrateV1ToV2,
+  useTabStore,
+} from "./tab-store";
+
+beforeEach(() => {
+  createTabRouterMock.mockClear();
+  useTabStore.getState().reset();
+});
+
+describe("sanitizeTabPath", () => {
+  it("rejects the root sentinel — tabs must be workspace-scoped", () => {
+    expect(sanitizeTabPath("/")).toBeNull();
+    expect(sanitizeTabPath("")).toBeNull();
+  });
+
+  it("silently rejects transition paths (no warn — navigation adapter intercepts them)", () => {
+    const warn = vi.spyOn(console, "warn").mockImplementation(() => {});
+    expect(sanitizeTabPath("/workspaces/new")).toBeNull();
+    expect(sanitizeTabPath("/invite/abc")).toBeNull();
+    expect(warn).not.toHaveBeenCalled();
+    warn.mockRestore();
+  });
+
+  it("passes through valid workspace-scoped paths", () => {
+    expect(sanitizeTabPath("/acme/issues")).toBe("/acme/issues");
+    expect(sanitizeTabPath("/my-team/projects/abc")).toBe("/my-team/projects/abc");
+  });
+
+  it("rejects paths whose first segment is a reserved slug (missing workspace prefix)", () => {
+    const warn = vi.spyOn(console, "warn").mockImplementation(() => {});
+    expect(sanitizeTabPath("/issues")).toBeNull();
+    expect(sanitizeTabPath("/settings")).toBeNull();
+    expect(warn).toHaveBeenCalled();
+    warn.mockRestore();
+  });
+
+  it("passes through user slugs that happen to look path-like but aren't reserved", () => {
+    expect(sanitizeTabPath("/acme-issues/issues")).toBe("/acme-issues/issues");
+    expect(sanitizeTabPath("/project-x/inbox")).toBe("/project-x/inbox");
+  });
+});
+
+describe("migrateV1ToV2", () => {
+  it("groups v1 flat tabs by workspace slug", () => {
+    const v1 = {
+      tabs: [
+        { id: "t1", path: "/acme/issues", title: "Issues", icon: "ListTodo" },
+        { id: "t2", path: "/acme/projects", title: "Projects", icon: "FolderKanban" },
+        { id: "t3", path: "/butter/issues", title: "Issues", icon: "ListTodo" },
+      ],
+      activeTabId: "t2",
+    };
+    const v2 = migrateV1ToV2(v1);
+    expect(Object.keys(v2.byWorkspace).sort()).toEqual(["acme", "butter"]);
+    expect(v2.byWorkspace.acme.tabs).toHaveLength(2);
+    expect(v2.byWorkspace.butter.tabs).toHaveLength(1);
+    expect(v2.byWorkspace.acme.activeTabId).toBe("t2");
+    expect(v2.byWorkspace.butter.activeTabId).toBe("t3"); // first tab in group
+    expect(v2.activeWorkspaceSlug).toBe("acme"); // contained v1.activeTabId
+  });
+
+  it("drops tabs at root / transition / reserved-slug paths", () => {
+    const v1 = {
+      tabs: [
+        { id: "t1", path: "/", title: "Issues", icon: "ListTodo" },
+        { id: "t2", path: "/workspaces/new", title: "New", icon: "Plus" },
+        { id: "t3", path: "/invite/abc", title: "Invite", icon: "Mail" },
+        { id: "t4", path: "/acme/issues", title: "Issues", icon: "ListTodo" },
+      ],
+      activeTabId: "t1",
+    };
+    const v2 = migrateV1ToV2(v1);
+    expect(Object.keys(v2.byWorkspace)).toEqual(["acme"]);
+    expect(v2.byWorkspace.acme.tabs).toHaveLength(1);
+    // v1.activeTabId was dropped; active falls back to first group's first tab.
+    expect(v2.activeWorkspaceSlug).toBe("acme");
+    expect(v2.byWorkspace.acme.activeTabId).toBe("t4");
+  });
+
+  it("handles empty v1 state gracefully", () => {
+    const v2 = migrateV1ToV2({ tabs: [], activeTabId: "" });
+    expect(v2.byWorkspace).toEqual({});
+    expect(v2.activeWorkspaceSlug).toBeNull();
+  });
+
+  it("handles v1 with no tabs field (corrupted state)", () => {
+    const v2 = migrateV1ToV2({});
+    expect(v2.byWorkspace).toEqual({});
+    expect(v2.activeWorkspaceSlug).toBeNull();
+  });
+});
+
+describe("useTabStore actions", () => {
+  it("switchWorkspace creates a new group with a default tab on first entry", () => {
+    useTabStore.getState().switchWorkspace("acme");
+    const s = useTabStore.getState();
+    expect(s.activeWorkspaceSlug).toBe("acme");
+    expect(s.byWorkspace.acme.tabs).toHaveLength(1);
+    expect(s.byWorkspace.acme.tabs[0].path).toBe("/acme/issues");
+  });
+
+  it("switchWorkspace without openPath restores the group's last active tab", () => {
+    const store = useTabStore.getState();
+    store.switchWorkspace("acme");
+    store.addTab("/acme/projects", "Projects", "FolderKanban");
+    const acmeProjectsId = useTabStore.getState().byWorkspace.acme.tabs[1].id;
+    store.setActiveTab(acmeProjectsId);
+
+    // Enter a different workspace then come back
+    store.switchWorkspace("butter");
+    expect(useTabStore.getState().activeWorkspaceSlug).toBe("butter");
+
+    store.switchWorkspace("acme");
+    const s = useTabStore.getState();
+    expect(s.activeWorkspaceSlug).toBe("acme");
+    expect(s.byWorkspace.acme.activeTabId).toBe(acmeProjectsId);
+  });
+
+  it("switchWorkspace with openPath dedupes into an existing tab with same path", () => {
+    const store = useTabStore.getState();
+    store.switchWorkspace("acme"); // creates default /acme/issues
+    store.addTab("/acme/projects", "Projects", "FolderKanban");
+
+    store.switchWorkspace("acme", "/acme/issues");
+    const s = useTabStore.getState();
+    expect(s.byWorkspace.acme.tabs).toHaveLength(2); // no duplicate created
+    const activeTab = s.byWorkspace.acme.tabs.find(
+      (t) => t.id === s.byWorkspace.acme.activeTabId,
+    );
+    expect(activeTab?.path).toBe("/acme/issues");
+  });
+
+  it("switchWorkspace with openPath not matching any tab adds a new tab", () => {
+    const store = useTabStore.getState();
+    store.switchWorkspace("acme");
+    store.switchWorkspace("acme", "/acme/issues/bug-42");
+    const s = useTabStore.getState();
+    expect(s.byWorkspace.acme.tabs).toHaveLength(2);
+    const activeTab = s.byWorkspace.acme.tabs.find(
+      (t) => t.id === s.byWorkspace.acme.activeTabId,
+    );
+    expect(activeTab?.path).toBe("/acme/issues/bug-42");
+  });
+
+  it("openTab dedupes by path within the active workspace", () => {
+    const store = useTabStore.getState();
+    store.switchWorkspace("acme");
+    const id1 = store.openTab("/acme/projects", "Projects", "FolderKanban");
+    const id2 = store.openTab("/acme/projects", "Projects", "FolderKanban");
+    expect(id1).toBe(id2);
+    expect(useTabStore.getState().byWorkspace.acme.tabs).toHaveLength(2); // default + projects
+  });
+
+  it("closeTab on the last tab in a workspace reseeds the default tab", () => {
+    const store = useTabStore.getState();
+    store.switchWorkspace("acme");
+    const onlyTabId = useTabStore.getState().byWorkspace.acme.tabs[0].id;
+    store.closeTab(onlyTabId);
+    const s = useTabStore.getState();
+    expect(s.byWorkspace.acme.tabs).toHaveLength(1);
+    expect(s.byWorkspace.acme.tabs[0].path).toBe("/acme/issues");
+    expect(s.byWorkspace.acme.tabs[0].id).not.toBe(onlyTabId); // fresh tab
+  });
+
+  it("validateWorkspaceSlugs drops groups for slugs not in the valid set and repoints active", () => {
+    const store = useTabStore.getState();
+    store.switchWorkspace("acme");
+    store.switchWorkspace("butter");
+    store.switchWorkspace("acme");
+    expect(useTabStore.getState().activeWorkspaceSlug).toBe("acme");
+
+    // Admin removed the user from acme
+    store.validateWorkspaceSlugs(new Set(["butter"]));
+    const s = useTabStore.getState();
+    expect(Object.keys(s.byWorkspace)).toEqual(["butter"]);
+    expect(s.activeWorkspaceSlug).toBe("butter");
+  });
+
+  it("validateWorkspaceSlugs sets activeWorkspaceSlug to null when all groups are dropped", () => {
+    const store = useTabStore.getState();
+    store.switchWorkspace("acme");
+    store.validateWorkspaceSlugs(new Set());
+    const s = useTabStore.getState();
+    expect(s.byWorkspace).toEqual({});
+    expect(s.activeWorkspaceSlug).toBeNull();
+  });
+
+  it("reset wipes the whole store", () => {
+    const store = useTabStore.getState();
+    store.switchWorkspace("acme");
+    store.switchWorkspace("butter");
+    store.reset();
+    const s = useTabStore.getState();
+    expect(s.activeWorkspaceSlug).toBeNull();
+    expect(s.byWorkspace).toEqual({});
+  });
+
+  it("setActiveTab across workspaces also flips the active workspace", () => {
+    const store = useTabStore.getState();
+    store.switchWorkspace("acme");
+    store.switchWorkspace("butter");
+    const acmeTabId = useTabStore.getState().byWorkspace.acme.tabs[0].id;
+    store.setActiveTab(acmeTabId);
+    expect(useTabStore.getState().activeWorkspaceSlug).toBe("acme");
+  });
+});

apps/desktop/src/renderer/src/stores/tab-store.ts

@@ -0,0 +1,705 @@
+import { create } from "zustand";
+import { createJSONStorage, persist } from "zustand/middleware";
+import { arrayMove } from "@dnd-kit/sortable";
+import { createPersistStorage, defaultStorage } from "@multica/core/platform";
+import { createSafeId } from "@multica/core/utils";
+import { isReservedSlug } from "@multica/core/paths";
+import type { DataRouter } from "react-router-dom";
+import { createTabRouter } from "../routes";
+
+// ---------------------------------------------------------------------------
+// Types
+// ---------------------------------------------------------------------------
+
+export interface Tab {
+  id: string;
+  /** Every tab path is workspace-scoped: `/{workspaceSlug}/{route}/...`. */
+  path: string;
+  title: string;
+  icon: string;
+  router: DataRouter;
+  historyIndex: number;
+  historyLength: number;
+}
+
+export interface WorkspaceTabGroup {
+  tabs: Tab[];
+  /** Must be a valid tab.id in `tabs`; the empty-tabs state is transient only. */
+  activeTabId: string;
+}
+
+interface TabStore {
+  /**
+   * The workspace currently visible in the TabBar / TabContent. Null in three
+   * cases:
+   *   - Fresh install, before any workspace exists or is selected.
+   *   - Logged-out state (reset() wipes it).
+   *   - Every workspace the user had access to got deleted / revoked.
+   * When null, TabContent renders nothing and the WindowOverlay takes over.
+   */
+  activeWorkspaceSlug: string | null;
+
+  /**
+   * Tab groups keyed by workspace slug. Each slug maps to an independent
+   * (tabs, activeTabId) pair; switching workspaces swaps the visible set
+   * without affecting any other group. Cross-workspace tab leakage — the
+   * bug that drove this refactor — is impossible by construction because
+   * there is no global tab array anymore.
+   */
+  byWorkspace: Record<string, WorkspaceTabGroup>;
+
+  /**
+   * Switch to a workspace.
+   *   - If the group doesn't exist yet, create it with a single default tab.
+   *   - If `openPath` is given, find a tab with that exact path and activate
+   *     it; otherwise add a new tab and activate it.
+   *   - If `openPath` is omitted, restore the group's last active tab
+   *     (VSCode / Slack behavior — workspaces resume where you left off).
+   */
+  switchWorkspace: (slug: string, openPath?: string) => void;
+  /** Open-or-activate (dedupes by path) a tab in the active workspace. */
+  openTab: (path: string, title: string, icon: string) => string;
+  /** Always creates a new tab (no dedupe) in the active workspace. */
+  addTab: (path: string, title: string, icon: string) => string;
+  /**
+   * Close a tab. Finds it across all workspaces (callers like the X button
+   * only know the tab id, not the owning workspace). If this is the last
+   * tab in its workspace, reseed a default tab so the invariant
+   * "every live workspace has at least one tab" holds.
+   */
+  closeTab: (tabId: string) => void;
+  /**
+   * Activate a tab. Finds it across all workspaces. Sets both the owning
+   * workspace as active and that group's activeTabId; needed for any code
+   * path that "jumps" to a tab belonging to a non-active workspace.
+   */
+  setActiveTab: (tabId: string) => void;
+  /** Patch metadata of a tab (router-sync, title-sync). Finds across groups. */
+  updateTab: (tabId: string, patch: Partial<Pick<Tab, "path" | "title" | "icon">>) => void;
+  /** Patch history tracking of a tab. Finds across groups. */
+  updateTabHistory: (tabId: string, historyIndex: number, historyLength: number) => void;
+  /** Reorder within the active workspace's group only. */
+  moveTab: (fromIndex: number, toIndex: number) => void;
+  /**
+   * After the workspace list arrives/changes (login, realtime delete), drop
+   * any tab group whose slug is no longer in `validSlugs`, and repoint
+   * `activeWorkspaceSlug` if it pointed at one of the dropped groups.
+   */
+  validateWorkspaceSlugs: (validSlugs: Set<string>) => void;
+  /**
+   * Wipe everything. Called from logout so the next user doesn't inherit
+   * the prior user's tabs. Zustand persist only writes to localStorage;
+   * clearing the storage key alone would leave this live store intact
+   * until app restart.
+   */
+  reset: () => void;
+}
+
+// ---------------------------------------------------------------------------
+// Route → icon mapping (title comes from document.title, not from here)
+// ---------------------------------------------------------------------------
+
+const ROUTE_ICONS: Record<string, string> = {
+  inbox: "Inbox",
+  "my-issues": "CircleUser",
+  issues: "ListTodo",
+  projects: "FolderKanban",
+  autopilots: "ListTodo",
+  agents: "Bot",
+  runtimes: "Monitor",
+  skills: "BookOpenText",
+  settings: "Settings",
+};
+
+/**
+ * Resolve a route icon from a pathname.
+ *
+ * Tab paths are always workspace-scoped: `/{slug}/{route}/...`, so the route
+ * segment lives at index 1. Pre-workspace flows (create, invite) are rendered
+ * by the window overlay, never as tabs.
+ *
+ * Title is NOT determined here — it comes from document.title.
+ */
+export function resolveRouteIcon(pathname: string): string {
+  const segments = pathname.split("/").filter(Boolean);
+  return ROUTE_ICONS[segments[1] ?? ""] ?? "ListTodo";
+}
+
+/** Extract the leading workspace slug from a path, or null if the path
+ *  isn't workspace-scoped (global path, root, or empty). */
+function extractWorkspaceSlug(path: string): string | null {
+  const first = path.split("/").filter(Boolean)[0] ?? "";
+  if (!first) return null;
+  if (isReservedSlug(first)) return null;
+  return first;
+}
+
+// ---------------------------------------------------------------------------
+// Path sanitization (defensive)
+// ---------------------------------------------------------------------------
+
+/**
+ * Defensive: catch paths that don't belong in the tab store.
+ *
+ * Two kinds of rejects:
+ *  1. **Transition paths** (`/workspaces/new`, `/invite/...`). These are
+ *     pre-workspace flows rendered by the window overlay on desktop, not
+ *     tab routes. The navigation adapter normally intercepts these before
+ *     they reach the store; this guard catches older persisted state.
+ *  2. **Malformed workspace-scoped paths** like a stray `/issues/abc` that
+ *     was constructed without the workspace prefix. The router would
+ *     interpret `issues` as a workspace slug → NoAccessPage.
+ *
+ * Returns null for rejects (caller decides how to recover — usually by
+ * dropping the tab or substituting a default). Unlike the prior design,
+ * there is no root "/" sentinel — tabs are always scoped.
+ */
+export function sanitizeTabPath(path: string): string | null {
+  const firstSegment = path.split("/").filter(Boolean)[0] ?? "";
+  if (!firstSegment) return null;
+  if (isReservedSlug(firstSegment)) {
+    // Don't log for known transition paths — these are legitimate inputs
+    // at the interception boundary (older persisted state or stale callers).
+    const isTransition = path === "/workspaces/new" || path.startsWith("/invite/");
+    if (!isTransition) {
+      // eslint-disable-next-line no-console
+      console.warn(
+        `[tab-store] tab path "${path}" starts with reserved slug "${firstSegment}" — ` +
+          `caller likely forgot the workspace prefix. Dropping.`,
+      );
+    }
+    return null;
+  }
+  return path;
+}
+
+// ---------------------------------------------------------------------------
+// Tab factory
+// ---------------------------------------------------------------------------
+
+function createId(): string {
+  return createSafeId();
+}
+
+function makeTab(path: string, title: string, icon: string): Tab {
+  return {
+    id: createId(),
+    path,
+    title,
+    icon,
+    router: createTabRouter(path),
+    historyIndex: 0,
+    historyLength: 1,
+  };
+}
+
+/** Default entry point for a workspace — its issues list. */
+function defaultPathFor(slug: string): string {
+  return `/${slug}/issues`;
+}
+
+function defaultTabFor(slug: string): Tab {
+  const path = defaultPathFor(slug);
+  return makeTab(path, "Issues", resolveRouteIcon(path));
+}
+
+// ---------------------------------------------------------------------------
+// Group helpers
+// ---------------------------------------------------------------------------
+
+function findTabLocation(
+  byWorkspace: Record<string, WorkspaceTabGroup>,
+  tabId: string,
+): { slug: string; group: WorkspaceTabGroup; index: number } | null {
+  for (const slug of Object.keys(byWorkspace)) {
+    const group = byWorkspace[slug];
+    const index = group.tabs.findIndex((t) => t.id === tabId);
+    if (index >= 0) return { slug, group, index };
+  }
+  return null;
+}
+
+// ---------------------------------------------------------------------------
+// Store
+// ---------------------------------------------------------------------------
+
+export const useTabStore = create<TabStore>()(
+  persist(
+    (set, get) => ({
+      activeWorkspaceSlug: null,
+      byWorkspace: {},
+
+      switchWorkspace(slug, openPath) {
+        // Defensive no-op if slug is empty/invalid — callers like the
+        // NavigationAdapter's path-parser should already have filtered
+        // these, but belt-and-braces keeps garbage out of the store.
+        if (!slug) return;
+        const { byWorkspace } = get();
+        const existing = byWorkspace[slug];
+
+        // Decide the desired active path for this workspace.
+        const desiredPath = openPath ?? (existing ? null : defaultPathFor(slug));
+
+        if (!existing) {
+          // First time entering this workspace — create the group.
+          const seedPath =
+            desiredPath && sanitizeTabPath(desiredPath) === desiredPath
+              ? desiredPath
+              : defaultPathFor(slug);
+          const tab = makeTab(seedPath, "Issues", resolveRouteIcon(seedPath));
+          set({
+            activeWorkspaceSlug: slug,
+            byWorkspace: {
+              ...byWorkspace,
+              [slug]: { tabs: [tab], activeTabId: tab.id },
+            },
+          });
+          return;
+        }
+
+        // Workspace already has tabs. Either dedupe into an existing tab or
+        // add a new one (when openPath was supplied and no tab matches it).
+        if (desiredPath) {
+          const clean = sanitizeTabPath(desiredPath);
+          if (clean) {
+            const match = existing.tabs.find((t) => t.path === clean);
+            if (match) {
+              set({
+                activeWorkspaceSlug: slug,
+                byWorkspace: {
+                  ...byWorkspace,
+                  [slug]: { ...existing, activeTabId: match.id },
+                },
+              });
+              return;
+            }
+            const tab = makeTab(clean, "Issues", resolveRouteIcon(clean));
+            set({
+              activeWorkspaceSlug: slug,
+              byWorkspace: {
+                ...byWorkspace,
+                [slug]: {
+                  tabs: [...existing.tabs, tab],
+                  activeTabId: tab.id,
+                },
+              },
+            });
+            return;
+          }
+        }
+
+        // No openPath (or openPath was rejected) — just restore the group.
+        set({ activeWorkspaceSlug: slug });
+      },
+
+      openTab(path, title, icon) {
+        const { activeWorkspaceSlug, byWorkspace } = get();
+        const clean = sanitizeTabPath(path);
+        if (!activeWorkspaceSlug || !clean) return "";
+        const group = byWorkspace[activeWorkspaceSlug];
+        if (!group) return "";
+
+        const existing = group.tabs.find((t) => t.path === clean);
+        if (existing) {
+          set({
+            byWorkspace: {
+              ...byWorkspace,
+              [activeWorkspaceSlug]: { ...group, activeTabId: existing.id },
+            },
+          });
+          return existing.id;
+        }
+
+        const tab = makeTab(clean, title, icon);
+        set({
+          byWorkspace: {
+            ...byWorkspace,
+            [activeWorkspaceSlug]: {
+              tabs: [...group.tabs, tab],
+              activeTabId: group.activeTabId,
+            },
+          },
+        });
+        return tab.id;
+      },
+
+      addTab(path, title, icon) {
+        const { activeWorkspaceSlug, byWorkspace } = get();
+        const clean = sanitizeTabPath(path);
+        if (!activeWorkspaceSlug || !clean) return "";
+        const group = byWorkspace[activeWorkspaceSlug];
+        if (!group) return "";
+
+        const tab = makeTab(clean, title, icon);
+        set({
+          byWorkspace: {
+            ...byWorkspace,
+            [activeWorkspaceSlug]: {
+              tabs: [...group.tabs, tab],
+              activeTabId: group.activeTabId,
+            },
+          },
+        });
+        return tab.id;
+      },
+
+      closeTab(tabId) {
+        const { byWorkspace } = get();
+        const hit = findTabLocation(byWorkspace, tabId);
+        if (!hit) return;
+        const { slug, group, index } = hit;
+
+        const closing = group.tabs[index];
+        closing.router.dispose();
+
+        if (group.tabs.length === 1) {
+          // Last tab in this workspace — reseed a default so the workspace
+          // always has at least one tab. Closing a workspace as an explicit
+          // action is a separate concern (Leave/Delete in Settings).
+          const fresh = defaultTabFor(slug);
+          set({
+            byWorkspace: {
+              ...byWorkspace,
+              [slug]: { tabs: [fresh], activeTabId: fresh.id },
+            },
+          });
+          return;
+        }
+
+        const nextTabs = group.tabs.filter((t) => t.id !== tabId);
+        const nextActiveTabId =
+          group.activeTabId === tabId
+            ? nextTabs[Math.min(index, nextTabs.length - 1)].id
+            : group.activeTabId;
+
+        set({
+          byWorkspace: {
+            ...byWorkspace,
+            [slug]: { tabs: nextTabs, activeTabId: nextActiveTabId },
+          },
+        });
+      },
+
+      setActiveTab(tabId) {
+        const { byWorkspace, activeWorkspaceSlug } = get();
+        const hit = findTabLocation(byWorkspace, tabId);
+        if (!hit) return;
+        const { slug, group } = hit;
+        if (slug === activeWorkspaceSlug && group.activeTabId === tabId) return;
+        set({
+          activeWorkspaceSlug: slug,
+          byWorkspace: {
+            ...byWorkspace,
+            [slug]: { ...group, activeTabId: tabId },
+          },
+        });
+      },
+
+      updateTab(tabId, patch) {
+        const { byWorkspace } = get();
+        const hit = findTabLocation(byWorkspace, tabId);
+        if (!hit) return;
+        const { slug, group, index } = hit;
+        const current = group.tabs[index];
+        const next: Tab = { ...current, ...patch };
+        const nextTabs = [...group.tabs];
+        nextTabs[index] = next;
+        set({
+          byWorkspace: {
+            ...byWorkspace,
+            [slug]: { ...group, tabs: nextTabs },
+          },
+        });
+      },
+
+      updateTabHistory(tabId, historyIndex, historyLength) {
+        const { byWorkspace } = get();
+        const hit = findTabLocation(byWorkspace, tabId);
+        if (!hit) return;
+        const { slug, group, index } = hit;
+        const current = group.tabs[index];
+        const next: Tab = { ...current, historyIndex, historyLength };
+        const nextTabs = [...group.tabs];
+        nextTabs[index] = next;
+        set({
+          byWorkspace: {
+            ...byWorkspace,
+            [slug]: { ...group, tabs: nextTabs },
+          },
+        });
+      },
+
+      moveTab(fromIndex, toIndex) {
+        if (fromIndex === toIndex) return;
+        const { activeWorkspaceSlug, byWorkspace } = get();
+        if (!activeWorkspaceSlug) return;
+        const group = byWorkspace[activeWorkspaceSlug];
+        if (!group) return;
+        set({
+          byWorkspace: {
+            ...byWorkspace,
+            [activeWorkspaceSlug]: {
+              ...group,
+              tabs: arrayMove(group.tabs, fromIndex, toIndex),
+            },
+          },
+        });
+      },
+
+      validateWorkspaceSlugs(validSlugs) {
+        const { activeWorkspaceSlug, byWorkspace } = get();
+        let changed = false;
+        const nextByWorkspace: Record<string, WorkspaceTabGroup> = {};
+        for (const slug of Object.keys(byWorkspace)) {
+          if (validSlugs.has(slug)) {
+            nextByWorkspace[slug] = byWorkspace[slug];
+          } else {
+            changed = true;
+            for (const t of byWorkspace[slug].tabs) t.router.dispose();
+          }
+        }
+
+        let nextActive = activeWorkspaceSlug;
+        if (nextActive && !validSlugs.has(nextActive)) {
+          nextActive = Object.keys(nextByWorkspace)[0] ?? null;
+          changed = true;
+        }
+
+        if (!changed) return;
+        set({ byWorkspace: nextByWorkspace, activeWorkspaceSlug: nextActive });
+      },
+
+      reset() {
+        const { byWorkspace } = get();
+        for (const slug of Object.keys(byWorkspace)) {
+          for (const t of byWorkspace[slug].tabs) t.router.dispose();
+        }
+        set({ activeWorkspaceSlug: null, byWorkspace: {} });
+      },
+    }),
+    {
+      name: "multica_tabs",
+      version: 2,
+      storage: createJSONStorage(() => createPersistStorage(defaultStorage)),
+      migrate: (persistedState, version) => {
+        // v1 → v2: flat `tabs` array → per-workspace grouping.
+        // Tabs whose path isn't workspace-scoped (root `/`, login, etc.)
+        // are dropped — they have no workspace to belong to, and the new
+        // model's invariant is "every tab lives in a workspace group".
+        if (version < 2 && persistedState && typeof persistedState === "object") {
+          return migrateV1ToV2(persistedState as Partial<V1Persisted>);
+        }
+        return persistedState as V2Persisted;
+      },
+      partialize: (state) => ({
+        activeWorkspaceSlug: state.activeWorkspaceSlug,
+        byWorkspace: Object.fromEntries(
+          Object.entries(state.byWorkspace).map(([slug, group]) => [
+            slug,
+            {
+              activeTabId: group.activeTabId,
+              tabs: group.tabs.map(
+                ({ router: _router, historyIndex: _hi, historyLength: _hl, ...rest }) =>
+                  rest,
+              ),
+            },
+          ]),
+        ),
+      }),
+      merge: (persistedState, currentState) => {
+        const persisted = persistedState as Partial<V2Persisted> | undefined;
+        if (!persisted?.byWorkspace) return currentState;
+
+        const byWorkspace: Record<string, WorkspaceTabGroup> = {};
+        for (const [slug, pGroup] of Object.entries(persisted.byWorkspace)) {
+          const tabs: Tab[] = [];
+          for (const pTab of pGroup.tabs) {
+            const clean = sanitizeTabPath(pTab.path);
+            // Persisted path may have come from a stale version or a
+            // manual edit. Drop rather than rewrite so we never silently
+            // put users on a path that doesn't match the group's slug.
+            if (!clean || extractWorkspaceSlug(clean) !== slug) {
+              // eslint-disable-next-line no-console
+              console.warn(
+                `[tab-store] dropping persisted tab "${pTab.path}" from ` +
+                  `group "${slug}" — path/slug mismatch`,
+              );
+              continue;
+            }
+            tabs.push({
+              id: pTab.id,
+              path: clean,
+              title: pTab.title,
+              icon: pTab.icon,
+              router: createTabRouter(clean),
+              historyIndex: 0,
+              historyLength: 1,
+            });
+          }
+          if (tabs.length === 0) continue;
+          const activeTabId = tabs.some((t) => t.id === pGroup.activeTabId)
+            ? pGroup.activeTabId
+            : tabs[0].id;
+          byWorkspace[slug] = { tabs, activeTabId };
+        }
+
+        const activeWorkspaceSlug =
+          persisted.activeWorkspaceSlug && byWorkspace[persisted.activeWorkspaceSlug]
+            ? persisted.activeWorkspaceSlug
+            : (Object.keys(byWorkspace)[0] ?? null);
+
+        return { ...currentState, byWorkspace, activeWorkspaceSlug };
+      },
+    },
+  ),
+);
+
+// ---------------------------------------------------------------------------
+// Persisted shapes (for migration)
+// ---------------------------------------------------------------------------
+
+interface V1Tab {
+  id: string;
+  path: string;
+  title: string;
+  icon: string;
+}
+
+interface V1Persisted {
+  tabs: V1Tab[];
+  activeTabId: string;
+}
+
+interface V2PersistedTab {
+  id: string;
+  path: string;
+  title: string;
+  icon: string;
+}
+
+interface V2PersistedGroup {
+  tabs: V2PersistedTab[];
+  activeTabId: string;
+}
+
+interface V2Persisted {
+  activeWorkspaceSlug: string | null;
+  byWorkspace: Record<string, V2PersistedGroup>;
+}
+
+export function migrateV1ToV2(v1: Partial<V1Persisted>): V2Persisted {
+  const byWorkspace: Record<string, V2PersistedGroup> = {};
+  const oldTabs = v1.tabs ?? [];
+  for (const tab of oldTabs) {
+    const slug = extractWorkspaceSlug(tab.path);
+    if (!slug) continue; // drop root / global-path tabs
+    if (!byWorkspace[slug]) byWorkspace[slug] = { tabs: [], activeTabId: "" };
+    byWorkspace[slug].tabs.push({
+      id: tab.id,
+      path: tab.path,
+      title: tab.title,
+      icon: tab.icon,
+    });
+  }
+
+  // Each group needs a valid activeTabId. Prefer the one from v1 if it
+  // landed in this group; otherwise fall back to the first tab.
+  for (const slug of Object.keys(byWorkspace)) {
+    const group = byWorkspace[slug];
+    const hasOldActive = group.tabs.some((t) => t.id === v1.activeTabId);
+    group.activeTabId = hasOldActive
+      ? (v1.activeTabId as string)
+      : group.tabs[0].id;
+  }
+
+  // Active workspace: whichever group inherited the v1 activeTab, falling
+  // back to the first group we created (arbitrary but deterministic given
+  // Object.keys iteration order on string keys).
+  let activeWorkspaceSlug: string | null = null;
+  for (const slug of Object.keys(byWorkspace)) {
+    if (byWorkspace[slug].activeTabId === v1.activeTabId) {
+      activeWorkspaceSlug = slug;
+      break;
+    }
+  }
+  if (!activeWorkspaceSlug) {
+    activeWorkspaceSlug = Object.keys(byWorkspace)[0] ?? null;
+  }
+
+  return { activeWorkspaceSlug, byWorkspace };
+}
+
+// ---------------------------------------------------------------------------
+// Selectors (convenience hooks)
+// ---------------------------------------------------------------------------
+
+/**
+ * Pure non-hook helper — useful from event handlers / effects that already
+ * need `.getState()`. For React subscriptions prefer the stable selectors
+ * below.
+ */
+export function getActiveTab(s: TabStore): Tab | null {
+  if (!s.activeWorkspaceSlug) return null;
+  const group = s.byWorkspace[s.activeWorkspaceSlug];
+  if (!group) return null;
+  return group.tabs.find((t) => t.id === group.activeTabId) ?? null;
+}
+
+/**
+ * The active workspace's tab group, or null when no workspace is active.
+ *
+ * Zustand compares selector returns with `Object.is`. Because `updateTab`
+ * /  `updateTabHistory` replace the group object on every router tick
+ * (immutable update), this selector returns a new reference on every
+ * router event — that's fine for TabBar which needs to observe tab-list
+ * changes, but don't use this selector from components that only care
+ * about one primitive (use `useActiveTabHistory` / `useActiveTabRouter`
+ * instead).
+ */
+export function useActiveGroup(): WorkspaceTabGroup | null {
+  return useTabStore((s) =>
+    s.activeWorkspaceSlug ? (s.byWorkspace[s.activeWorkspaceSlug] ?? null) : null,
+  );
+}
+
+/**
+ * Active tab id + active workspace slug as a compact pair. Both primitives
+ * are stable across unrelated store updates — e.g. an inactive tab's
+ * router tick doesn't churn these, so consumers don't re-render.
+ *
+ * Useful anywhere you'd previously have reached for `useActiveTab()` and
+ * only needed the identity (for memoization, effect deps, ipc).
+ */
+export function useActiveTabIdentity(): { slug: string | null; tabId: string | null } {
+  const slug = useTabStore((s) => s.activeWorkspaceSlug);
+  const tabId = useTabStore((s) =>
+    s.activeWorkspaceSlug
+      ? (s.byWorkspace[s.activeWorkspaceSlug]?.activeTabId ?? null)
+      : null,
+  );
+  return { slug, tabId };
+}
+
+/**
+ * Active tab's router — a stable reference across tab updates, because
+ * routers are created once per tab and never replaced by `updateTab`.
+ * Subscribers only re-render when the active tab *changes*, not on
+ * router events within the current tab.
+ */
+export function useActiveTabRouter(): DataRouter | null {
+  return useTabStore((s) => getActiveTab(s)?.router ?? null);
+}
+
+/**
+ * History tracking for the active tab as primitives. Subscribers re-render
+ * only when the numeric index / length change (i.e. on actual navigations),
+ * not on unrelated store updates.
+ */
+export function useActiveTabHistory(): {
+  historyIndex: number;
+  historyLength: number;
+} {
+  const historyIndex = useTabStore((s) => getActiveTab(s)?.historyIndex ?? 0);
+  const historyLength = useTabStore((s) => getActiveTab(s)?.historyLength ?? 1);
+  return { historyIndex, historyLength };
+}

apps/desktop/src/renderer/src/stores/window-overlay-store.ts

@@ -0,0 +1,31 @@
+import { create } from "zustand";
+
+/**
+ * Window-level transition overlay: pre-workspace flows that are NOT pages
+ * inside a tab. Triggered by navigation-adapter interception, zero-workspace
+ * auto-redirect, or deep link; rendered above the tab system as a full-window
+ * takeover.
+ *
+ * These flows used to be routes (`/workspaces/new`, `/invite/:id`) but on
+ * desktop the URL is invisible to users — routes are an implementation detail
+ * of the tab system. Representing transitions as routes meant tabs tried to
+ * persist them, TabBar rendered on top, and invite deep-linking had no clean
+ * dispatch target. Modeling them as application state removes all three.
+ */
+export type WindowOverlay =
+  | { type: "new-workspace" }
+  | { type: "invite"; invitationId: string }
+  | { type: "invitations" }
+  | { type: "onboarding" };
+
+interface WindowOverlayStore {
+  overlay: WindowOverlay | null;
+  open: (overlay: WindowOverlay) => void;
+  close: () => void;
+}
+
+export const useWindowOverlayStore = create<WindowOverlayStore>((set) => ({
+  overlay: null,
+  open: (overlay) => set({ overlay }),
+  close: () => set({ overlay: null }),
+}));

apps/desktop/src/shared/daemon-types.ts

@@ -0,0 +1,85 @@
+export type DaemonState =
+  | "running"
+  | "stopped"
+  | "starting"
+  | "stopping"
+  | "installing_cli"
+  | "cli_not_found";
+
+export interface DaemonStatus {
+  state: DaemonState;
+  pid?: number;
+  uptime?: string;
+  daemonId?: string;
+  deviceName?: string;
+  agents?: string[];
+  workspaceCount?: number;
+  /** CLI profile this daemon belongs to. Empty string means the default profile. */
+  profile?: string;
+  /** Backend URL the daemon connects to. */
+  serverUrl?: string;
+}
+
+export interface DaemonPrefs {
+  autoStart: boolean;
+  autoStop: boolean;
+}
+
+export const DAEMON_STATE_COLORS: Record<DaemonState, string> = {
+  running: "bg-emerald-500",
+  stopped: "bg-muted-foreground/40",
+  starting: "bg-amber-500 animate-pulse",
+  stopping: "bg-amber-500 animate-pulse",
+  installing_cli: "bg-sky-500 animate-pulse",
+  cli_not_found: "bg-red-500",
+};
+
+export const DAEMON_STATE_LABELS: Record<DaemonState, string> = {
+  running: "Running",
+  stopped: "Stopped",
+  starting: "Starting…",
+  stopping: "Stopping…",
+  installing_cli: "Setting up…",
+  cli_not_found: "Setup Failed",
+};
+
+export function formatUptime(uptime?: string): string {
+  if (!uptime) return "";
+  const match = uptime.match(/(?:(\d+)h)?(\d+)m/);
+  if (!match) return uptime;
+  const h = match[1] ? `${match[1]}h ` : "";
+  const m = match[2] ? `${match[2]}m` : "";
+  return `${h}${m}`.trim() || uptime;
+}
+
+/**
+ * User-facing description for the local daemon's current state. Replaces the
+ * raw state label ("Running" / "Stopped") with a sentence that answers
+ * "what does this mean for me?" — i.e. whether tasks can run on this device.
+ *
+ * `runtimeCount` is the number of runtimes the local daemon has registered
+ * (claude / codex / gemini / ... — one per detected CLI). It's only consulted
+ * when state === "running".
+ */
+export function daemonStateDescription(state: DaemonState, runtimeCount: number): string {
+  switch (state) {
+    case "running":
+      if (runtimeCount === 0) {
+        return "Running, but no runtimes have registered yet.";
+      }
+      if (runtimeCount === 1) {
+        return "Running here · 1 runtime available for tasks.";
+      }
+      return `Running here · ${runtimeCount} runtimes available for tasks.`;
+    case "stopped":
+      return "Not running · this device can't take new tasks.";
+    case "starting":
+      return "Starting up the local daemon…";
+    case "stopping":
+      return "Shutting down the local daemon…";
+    case "installing_cli":
+      return "Setting up the runtime for the first time. Only happens once.";
+    case "cli_not_found":
+      return "Setup failed · couldn't download the runtime. Check your network.";
+  }
+}

apps/desktop/src/shared/runtime-config.test.ts

@@ -0,0 +1,151 @@
+import { describe, expect, it } from "vitest";
+import {
+  DEFAULT_RUNTIME_CONFIG,
+  deriveWsUrl,
+  parseRuntimeConfig,
+  runtimeConfigFromDevEnv,
+} from "./runtime-config";
+
+describe("runtime config", () => {
+  it("uses cloud defaults without a desktop.json file", () => {
+    expect(DEFAULT_RUNTIME_CONFIG).toEqual({
+      schemaVersion: 1,
+      apiUrl: "https://api.multica.ai",
+      wsUrl: "wss://api.multica.ai/ws",
+      appUrl: "https://multica.ai",
+    });
+  });
+
+  it("derives https/wss compatible URLs from apiUrl", () => {
+    expect(
+      parseRuntimeConfig(
+        JSON.stringify({
+          schemaVersion: 1,
+          apiUrl: "https://congvc-x99.taila6fa8a.ts.net:18443",
+        }),
+      ),
+    ).toEqual({
+      schemaVersion: 1,
+      apiUrl: "https://congvc-x99.taila6fa8a.ts.net:18443",
+      wsUrl: "wss://congvc-x99.taila6fa8a.ts.net:18443/ws",
+      appUrl: "https://congvc-x99.taila6fa8a.ts.net:18443",
+    });
+  });
+
+  it("strips the leading api. label when deriving appUrl", () => {
+    expect(
+      parseRuntimeConfig(
+        JSON.stringify({ schemaVersion: 1, apiUrl: "https://api.multica.ai" }),
+      ),
+    ).toEqual({
+      schemaVersion: 1,
+      apiUrl: "https://api.multica.ai",
+      wsUrl: "wss://api.multica.ai/ws",
+      appUrl: "https://multica.ai",
+    });
+  });
+
+  it("derives ws for http api URLs", () => {
+    expect(deriveWsUrl("http://localhost:8080")).toBe("ws://localhost:8080/ws");
+  });
+
+  it("accepts explicit appUrl and wsUrl", () => {
+    expect(
+      parseRuntimeConfig(
+        JSON.stringify({
+          schemaVersion: 1,
+          apiUrl: "https://api.example.com/",
+          wsUrl: "wss://ws.example.com/socket/",
+          appUrl: "https://app.example.com/",
+        }),
+      ),
+    ).toEqual({
+      schemaVersion: 1,
+      apiUrl: "https://api.example.com",
+      wsUrl: "wss://ws.example.com/socket",
+      appUrl: "https://app.example.com",
+    });
+  });
+
+  it("rejects invalid JSON", () => {
+    expect(() => parseRuntimeConfig("{")).toThrow(/Invalid desktop runtime config JSON/);
+  });
+
+  it("rejects unsupported schema versions", () => {
+    expect(() =>
+      parseRuntimeConfig(JSON.stringify({ schemaVersion: 2, apiUrl: "https://api.example.com" })),
+    ).toThrow(/schemaVersion/);
+  });
+
+  it("rejects non-http api schemes", () => {
+    expect(() =>
+      parseRuntimeConfig(JSON.stringify({ schemaVersion: 1, apiUrl: "file:///tmp/multica" })),
+    ).toThrow(/apiUrl must use http or https/);
+  });
+
+  it("rejects non-ws websocket schemes", () => {
+    expect(() =>
+      parseRuntimeConfig(
+        JSON.stringify({
+          schemaVersion: 1,
+          apiUrl: "https://api.example.com",
+          wsUrl: "https://api.example.com/ws",
+        }),
+      ),
+    ).toThrow(/wsUrl must use ws or wss/);
+  });
+
+  it("preserves electron-vite dev env precedence", () => {
+    expect(
+      runtimeConfigFromDevEnv({
+        apiUrl: "http://dev-api.example.test:8080/",
+        wsUrl: "ws://dev-api.example.test:8080/ws/",
+        appUrl: "http://dev-app.example.test:3000/",
+      }),
+    ).toEqual({
+      schemaVersion: 1,
+      apiUrl: "http://dev-api.example.test:8080",
+      wsUrl: "ws://dev-api.example.test:8080/ws",
+      appUrl: "http://dev-app.example.test:3000",
+    });
+  });
+
+  it("falls back to local web URL when dev apiUrl is localhost", () => {
+    expect(runtimeConfigFromDevEnv({ apiUrl: "http://localhost:8080" })).toEqual({
+      schemaVersion: 1,
+      apiUrl: "http://localhost:8080",
+      wsUrl: "ws://localhost:8080/ws",
+      appUrl: "http://localhost:3000",
+    });
+  });
+
+  it("derives dev appUrl by stripping the leading api. label", () => {
+    // When the dev renderer is pointed at a remote backend (e.g. a test
+    // environment), copy-link / share URLs must reflect that environment's
+    // public web host, not the api host. Multica's convention exposes the
+    // api at `api.<web-host>`, so stripping the leading label gives the
+    // right web origin without a separate VITE_APP_URL.
+    expect(
+      runtimeConfigFromDevEnv({ apiUrl: "https://api.test.multica.ai" }),
+    ).toEqual({
+      schemaVersion: 1,
+      apiUrl: "https://api.test.multica.ai",
+      wsUrl: "wss://api.test.multica.ai/ws",
+      appUrl: "https://test.multica.ai",
+    });
+  });
+
+  it("dev VITE_APP_URL still wins over apiUrl-derived value", () => {
+    expect(
+      runtimeConfigFromDevEnv({
+        apiUrl: "https://api.test.multica.ai",
+        appUrl: "https://staging.multica.ai",
+      }),
+    ).toEqual({
+      schemaVersion: 1,
+      apiUrl: "https://api.test.multica.ai",
+      wsUrl: "wss://api.test.multica.ai/ws",
+      appUrl: "https://staging.multica.ai",
+    });
+  });
+});

apps/desktop/src/shared/runtime-config.ts

@@ -0,0 +1,179 @@
+export interface RuntimeConfig {
+  schemaVersion: 1;
+  apiUrl: string;
+  wsUrl: string;
+  appUrl: string;
+}
+
+export interface RuntimeConfigError {
+  message: string;
+}
+
+export type RuntimeConfigResult =
+  | { ok: true; config: RuntimeConfig }
+  | { ok: false; error: RuntimeConfigError };
+
+export const DEFAULT_RUNTIME_CONFIG: RuntimeConfig = Object.freeze({
+  schemaVersion: 1,
+  apiUrl: "https://api.multica.ai",
+  wsUrl: "wss://api.multica.ai/ws",
+  appUrl: "https://multica.ai",
+});
+
+const LOCAL_DEV_RUNTIME_CONFIG: RuntimeConfig = Object.freeze({
+  schemaVersion: 1,
+  apiUrl: "http://localhost:8080",
+  wsUrl: "ws://localhost:8080/ws",
+  appUrl: "http://localhost:3000",
+});
+
+export interface RuntimeConfigEnv {
+  apiUrl?: string;
+  wsUrl?: string;
+  appUrl?: string;
+}
+
+export function runtimeConfigFromDevEnv(env: RuntimeConfigEnv): RuntimeConfig {
+  const apiUrl = normalizeHttpUrl(
+    env.apiUrl || LOCAL_DEV_RUNTIME_CONFIG.apiUrl,
+    "VITE_API_URL",
+  );
+  return {
+    schemaVersion: 1,
+    apiUrl,
+    wsUrl: env.wsUrl
+      ? normalizeWsUrl(env.wsUrl, "VITE_WS_URL")
+      : deriveWsUrl(apiUrl),
+    appUrl: env.appUrl
+      ? normalizeHttpUrl(env.appUrl, "VITE_APP_URL")
+      : deriveDevAppUrl(apiUrl),
+  };
+}
+
+export function parseRuntimeConfig(raw: string): RuntimeConfig {
+  let parsed: unknown;
+  try {
+    parsed = JSON.parse(raw);
+  } catch (err) {
+    throw new Error(
+      `Invalid desktop runtime config JSON: ${err instanceof Error ? err.message : "parse failed"}`,
+    );
+  }
+
+  if (!parsed || typeof parsed !== "object" || Array.isArray(parsed)) {
+    throw new Error("Invalid desktop runtime config: expected a JSON object");
+  }
+
+  const obj = parsed as Record<string, unknown>;
+  if (obj.schemaVersion !== 1) {
+    throw new Error("Unsupported desktop runtime config schemaVersion: expected 1");
+  }
+
+  const apiUrl = requiredString(obj.apiUrl, "apiUrl");
+  const appUrl = optionalString(obj.appUrl, "appUrl");
+  const wsUrl = optionalString(obj.wsUrl, "wsUrl");
+
+  const normalizedApiUrl = normalizeHttpUrl(apiUrl, "apiUrl");
+  return {
+    schemaVersion: 1,
+    apiUrl: normalizedApiUrl,
+    wsUrl: wsUrl ? normalizeWsUrl(wsUrl, "wsUrl") : deriveWsUrl(normalizedApiUrl),
+    appUrl: appUrl ? normalizeHttpUrl(appUrl, "appUrl") : deriveAppUrl(normalizedApiUrl),
+  };
+}
+
+export function deriveWsUrl(apiUrl: string): string {
+  const url = new URL(apiUrl);
+  if (url.protocol === "https:") url.protocol = "wss:";
+  else if (url.protocol === "http:") url.protocol = "ws:";
+  else throw new Error("apiUrl must use http or https");
+  url.pathname = joinPath(url.pathname, "/ws");
+  url.search = "";
+  url.hash = "";
+  return trimTrailingSlash(url.toString());
+}
+
+// Convention: api hosts are exposed at `api.<web-host>` (api.multica.ai →
+// multica.ai, api.test.multica.ai → test.multica.ai). Strip the leading
+// `api.` label so a single `apiUrl` configuration produces the right
+// shareable web URL. Hosts that don't match the convention (no leading
+// `api.` label, or short two-label hosts like `api.local`) fall through
+// untouched — those deployments must set `appUrl` explicitly.
+export function deriveAppUrl(apiUrl: string): string {
+  const url = new URL(apiUrl);
+  url.pathname = "";
+  url.search = "";
+  url.hash = "";
+  if (url.hostname.startsWith("api.") && url.hostname.split(".").length >= 3) {
+    url.hostname = url.hostname.slice("api.".length);
+  }
+  return trimTrailingSlash(url.toString());
+}
+
+// Dev variant: when the api host is the local backend (`localhost:8080` /
+// `127.0.0.1:8080`), the renderer is served from a different port (3000),
+// so deriving by host alone is wrong. Fall back to the local dev web URL
+// in that case; for any non-local host (e.g. a remote test environment),
+// trust the production-style derivation so `apiUrl=https://api.test.x`
+// yields `appUrl=https://test.x` without a separate VITE_APP_URL.
+export function deriveDevAppUrl(apiUrl: string): string {
+  const url = new URL(apiUrl);
+  if (url.hostname === "localhost" || url.hostname === "127.0.0.1") {
+    return LOCAL_DEV_RUNTIME_CONFIG.appUrl;
+  }
+  return deriveAppUrl(apiUrl);
+}
+
+function requiredString(value: unknown, field: string): string {
+  if (typeof value !== "string" || value.trim().length === 0) {
+    throw new Error(`Invalid desktop runtime config: ${field} must be a non-empty string`);
+  }
+  return value;
+}
+
+function optionalString(value: unknown, field: string): string | undefined {
+  if (value === undefined) return undefined;
+  if (typeof value !== "string" || value.trim().length === 0) {
+    throw new Error(`Invalid desktop runtime config: ${field} must be a non-empty string when set`);
+  }
+  return value;
+}
+
+function normalizeHttpUrl(value: string, field: string): string {
+  let url: URL;
+  try {
+    url = new URL(value.trim());
+  } catch {
+    throw new Error(`Invalid desktop runtime config: ${field} must be a valid URL`);
+  }
+  if (url.protocol !== "http:" && url.protocol !== "https:") {
+    throw new Error(`Invalid desktop runtime config: ${field} must use http or https`);
+  }
+  url.search = "";
+  url.hash = "";
+  return trimTrailingSlash(url.toString());
+}
+
+function normalizeWsUrl(value: string, field: string): string {
+  let url: URL;
+  try {
+    url = new URL(value.trim());
+  } catch {
+    throw new Error(`Invalid desktop runtime config: ${field} must be a valid URL`);
+  }
+  if (url.protocol !== "ws:" && url.protocol !== "wss:") {
+    throw new Error(`Invalid desktop runtime config: ${field} must use ws or wss`);
+  }
+  url.search = "";
+  url.hash = "";
+  return trimTrailingSlash(url.toString());
+}
+
+function joinPath(base: string, suffix: string): string {
+  const normalizedBase = base.endsWith("/") ? base.slice(0, -1) : base;
+  return `${normalizedBase}${suffix}`;
+}
+
+function trimTrailingSlash(value: string): string {
+  return value.replace(/\/+$/, "");
+}

apps/desktop/test/setup.ts

@@ -0,0 +1,38 @@
+import "@testing-library/jest-dom/vitest";
+
+function createMemoryStorage(): Storage {
+  const values = new Map<string, string>();
+
+  return {
+    get length() {
+      return values.size;
+    },
+    clear: () => values.clear(),
+    getItem: (key: string) => values.get(key) ?? null,
+    key: (index: number) => Array.from(values.keys())[index] ?? null,
+    removeItem: (key: string) => {
+      values.delete(key);
+    },
+    setItem: (key: string, value: string) => {
+      values.set(key, value);
+    },
+  };
+}
+
+const localStorageIsUsable =
+  typeof globalThis.localStorage?.getItem === "function" &&
+  typeof globalThis.localStorage?.setItem === "function" &&
+  typeof globalThis.localStorage?.removeItem === "function" &&
+  typeof globalThis.localStorage?.clear === "function";
+
+if (!localStorageIsUsable) {
+  const storage = createMemoryStorage();
+  Object.defineProperty(globalThis, "localStorage", {
+    configurable: true,
+    value: storage,
+  });
+  Object.defineProperty(window, "localStorage", {
+    configurable: true,
+    value: storage,
+  });
+}