mirror of
https://github.com/danswer-ai/danswer.git
synced 2025-05-08 02:40:32 +02:00
* Initial backend changes for file connector * Add another background job to clean up files * UI + tweaks for backend
This serves as an example for how to deploy everything on a single machine. This is not optimal, but can get you started easily and cheaply. To run:
-
Run one of the docker compose commands below depending on your environment:
- For Local:
docker compose -f docker-compose.dev.yml -p danswer-stack up -d --build
- This will start Web/API servers, Postgres (backend DB), Qdrant (vector DB), and the background indexing job.
- Downloading packages/requirements may take 20+ minutes depending on your internet connection.
- For Local:
-
To shut down the deployment run (use stop to stop containers, down to remove containers):
- For Local:
docker compose -f docker-compose.dev.yml -p danswer-stack stop
- For Local:
-
To completely remove Danswer (WARNING, this will also erase your indexed data and all users) run:
- For Local:
docker compose -f docker-compose.dev.yml -p danswer-stack down -v
- For Local:
Additional steps for setting up for Prod:
-
Set up a
.env
file in this directory with relevant environment variables.- Refer to env.dev.template and env.prod.template
-
Set up https:
- Set up a
.env.nginx
file in this directory based onenv.nginx.template
. chmod +x init-letsencrypt.sh
+./init-letsencrypt.sh
to set up https certificate.
- Set up a
-
Follow the above steps but replacing dev with prod.