This is a project inspired by nof1 Alpha Arena, you can setup AI trading bot on crypto market.

DONE:

• Paper Trading
• OpenAI compatible API
• AWS Bedrock integration (Claude Sonnet 4.5, Qwen 3, DeepSeek 3.1)
• LEVERAGE
• cctx quotes

TODO:

• real trading

For detailed system architecture documentation, please refer to: architecture.md

This document includes:

• 🏗️ Overall system architecture design
• 🔄 Core business process descriptions
• 📊 Database design and model relationships
• 🔌 API interface documentation
• 🤖 AI decision service implementation details
• 📈 Trade execution and order matching mechanisms

Complete AWS Bedrock integration guide: AI Model Guidance.md

This document includes:

• 🎯 Supported AI models (Claude Sonnet 4.5, Qwen 3, DeepSeek 3.1)
• ⚙️ Detailed configuration steps and account setup
• 🔐 API Key acquisition and authentication methods
• 📊 API call chain sequence diagrams
• ❓ Frequently Asked Questions (FAQ)
• 🔧 Troubleshooting guide

• Node.js 18+ and pnpm
• Python 3.10+ and uv

# install JS deps and sync Python env
pnpm run install:all

By default, the workspace scripts launch:

• Backend on port 5611
• Frontend on port 5621

Start both dev servers:

pnpm run dev

Open:

• Frontend: http://localhost:5621
• Backend WS: ws://localhost:5611/ws

Important: The frontend source is currently configured for port 5621. To use the workspace defaults (5611), update the following in frontend/app/main.tsx:

• WebSocket URL: ws://localhost:5611/ws
• API_BASE: http://127.0.0.1:5611

Alternatively, run the backend on 5621:

# from repo root
cd backend
uv sync
uv run uvicorn main:app --reload --port  5621 --host 0.0.0.0

# build frontend; backend has no dedicated build step
pnpm run build

Static assets for the frontend are produced by Vite. The backend is a standard FastAPI app that can be run with Uvicorn or any ASGI server.

Referenced from another git repo: https://gist.github.com/wquguru/7d268099b8c04b7e5b6ad6fae922ae83

MIT