DEPLOY_RENDER.md

Deploy U2INVEST To Render

This repo is now prepared to run as a single Render web service:

• Flask API
• Built React frontend
• Video assets
• Persistent storage for SQLite, Chroma, and model cache

What Is Already Prepared

• Dockerfile builds the React frontend and serves the full app with Flask + Gunicorn.
• render.yaml defines a Render web service with a persistent disk and health check.
• Runtime paths are configurable through environment variables.
• The knowledge base retriever now initializes lazily so the app starts faster.

What You Need To Do

1. Push the latest repo state to GitHub.
2. Create a Render account and connect your GitHub account.
3. In Render, create a new Blueprint or Web Service from this repository.
4. If you use the Blueprint flow, Render will read render.yaml automatically.
5. When prompted for secrets, set:
   • DEEPSEEK_API_KEY
   • optionally replace FLASK_SECRET_KEY if you do not want the generated value
6. Confirm the persistent disk mount path is /app/data.
7. Deploy.
8. After deploy finishes, open:
   • /api/health
   • /
   • /app/chat

If Render Asks For A Payment Method

Use render-free.yaml instead of render.yaml.

That free option removes the persistent disk and changes the service to a free instance so you can test the full app without adding a card. See DEPLOY_RENDER_FREE.md for the exact steps.

Recommended First Production Check

Verify these items before changing the agent implementation:

• The home page loads from the Render URL.
• GET /api/health returns "status": "ok".
• The contact page works.
• The Trading Lab loads.
• U2CHAT returns a response when DEEPSEEK_API_KEY is set.

Local Docker Test

1. Copy .env.example to .env.
2. Fill in FLASK_SECRET_KEY and DEEPSEEK_API_KEY.
3. Run:

docker compose up --build

4. Open:

http://localhost:5000
http://localhost:5000/api/health

After The Site Is Live

Do this next:

1. Add your custom domain in Render.
2. Verify DNS and HTTPS.
3. Test the app again on the custom domain.
4. Only after the current backend is stable, start replacing the current agent with TradingAgents behind the existing /api/agent/* routes.

TradingAgents Migration Strategy

Keep the public API stable and swap the backend implementation in phases:

1. Keep the current frontend unchanged.
2. Keep /api/agent/chat, /api/agent/sessions, and /api/agent/history.
3. Replace the internals of agent_graph.py with a TradingAgents-backed adapter.
4. Compare latency, token cost, and output quality before fully switching.