Deployment

This page is a starter scaffold. Fill in production deployment details (Docker, reverse proxy, TLS) as your setup matures.

Running the Stack

Backend

uv sync --dev
uvicorn src.main:create_app --host 0.0.0.0 --port 8000

Or as a module:

python -m src.main

Frontend

Build the SPA and serve the static output, or serve it via your reverse proxy:

cd frontend
npm install
npm run build
# output in frontend/dist/

Configuration in Production

Set the production values for the variables described in Configuration, especially:

• LLM_API_KEY / LLM_BASE_URL / LLM_MODEL
• DATABASE_URL (point at your production database)
• SMTP_* for real invitation emails

Reverse Proxy Notes

• The API is served under /api/v1/.
• Chat uses SSE streaming — ensure your proxy does not buffer responses (disable proxy buffering for the chat route).
• Serve the built frontend static files and proxy /api/v1/ to the backend.

Future Work

• Container image / Docker Compose
• Managed PostgreSQL
• OIDC SSO wiring (currently a stub)