A modular multi-agent AI platform with specialized bots for coding, research, trading analysis, news, and generative AI — accessible via Telegram and a React web chat UI.
| Command | Bot | Description |
|---|---|---|
/code |
Coding Bot | AI code generation, review, refactoring — with git push + auto-deploy |
/research |
Research Bot | Web research and summaries with citations |
/trade |
Trading Bot | Market analysis and technical commentary |
/gen |
GenAI Bot | Creative writing, image prompts, style transfer |
/news |
News Bot | Financial news digest (also runs every 30 min automatically) |
/cron |
Cron Bot | Schedule recurring tasks (owner only) |
- Backend — Render (Docker, free tier)
- Database — Supabase (Postgres, free tier)
- Frontend — Vercel (Vite/React, free tier)
- LLM — Groq (free tier, OpenAI-compatible)
- Redis — not required, in-memory fallback used
See deployment.md for the full step-by-step guide.
# 1. Copy and fill in environment variables
cp .env.example .env
# 2. Start infrastructure (PostgreSQL — Redis and Ollama optional)
docker compose up postgres -d
# 3. Install Python dependencies
pip install -e ".[test]"
# 4. Run the backend
uvicorn backend.main:app --reload --port 8000
# 5. Run the frontend (separate terminal)
cd frontend && npm install && npm run devFor local LLM support, start Ollama separately and set OLLAMA_BASE_URL=http://localhost:11434 in .env.
For production, set OPENAI_BASE_URL=https://api.groq.com/openai/v1 and use your Groq API key.
pytest tests/unit -vdocker compose up --buildServices: backend (port 8000), frontend (port 3000), postgres.
Copy .env.example to .env and fill in your values. Key variables:
| Variable | Description |
|---|---|
DATABASE_URL |
Postgres connection string (Supabase in production) |
OPENAI_API_KEY |
Groq API key (or OpenAI) |
OPENAI_BASE_URL |
https://api.groq.com/openai/v1 for Groq |
TELEGRAM_TOKEN |
From @BotFather |
GIT_REMOTE |
Token-embedded GitHub URL for coding bot auto-deploy |
REDIS_URL |
Optional — omit to use in-memory fallback |