docs/DEPLOYMENT.md

Deployment Guide

Streamlit Community Cloud

1. Push the repository to GitHub.
2. Create a Streamlit app with streamlit_app.py as the entrypoint.
3. Use Python 3.12.
4. Add GEMINI_API_KEY only if optional LLM narration is desired.
5. Keep the default SQLite/local artifact mode for a stateless public demo.

The app is fully functional without an LLM key.

Render API

render.yaml defines a Docker-backed FastAPI service. Configure:

• DATABASE_URL: Neon, Supabase or another PostgreSQL URL.
• GEMINI_API_KEY: optional.
• persistent/object storage if generated artifacts must survive redeployments.

Docker Compose

docker compose up --build

• Streamlit: http://localhost:8501
• FastAPI: http://localhost:8000/docs

The isolated worker has no exposed port and no network.

Production topology

For heavier datasets, make /v1/analyze/* enqueue jobs to a separate worker and return 202 Accepted. Persist state in PostgreSQL, place artifacts in S3-compatible storage, and stream progress through server-sent events or polling.