Lean All-in-One

English | 简体中文 | 繁體中文 | 日本語

🚀 Lean Engine All-in-One

Docker License GPU

All-in-One Docker image for QuantConnect Lean algorithmic trading engine with GPU support, modern Web UI, REST API, and MCP integration.

✨ Features

• 🎯 GPU Auto-Selection - Automatically selects GPU with lowest memory usage
• 🌐 Modern Web UI - Responsive design with dark mode and i18n (EN/ZH/JA)
• 🔌 REST API - FastAPI with Swagger docs at /docs
• 🤖 MCP Server - Model Context Protocol for AI integration
• 📊 Real-time Monitoring - GPU status and backtest progress via WebSocket
• 🐳 All-in-One - Single container with nginx, API, and MCP services

🚀 Quick Start

docker run -d --gpus all \
  -p 8280:80 -p 8281:8000 -p 8282:8001 \
  --name lean-engine \
  neosun/lean:latest

Access:

• Web UI: http://localhost:8280
• API Docs: http://localhost:8281/docs
• Health: http://localhost:8281/health

📦 Installation

Docker Compose (Recommended)

services:
  lean:
    image: neosun/lean:latest
    container_name: lean-engine
    restart: unless-stopped
    ports:
      - "8280:80"      # Web UI
      - "8281:8000"    # REST API
      - "8282:8001"    # MCP Server
    volumes:
      - ./data:/Lean/Data
      - ./results:/Lean/Results
      - ./algorithms:/Lean/Algorithms
    deploy:
      resources:
        reservations:
          devices:
            - driver: nvidia
              count: all
              capabilities: [gpu]

docker compose up -d

⚙️ Configuration

Environment Variable	Default	Description
MODE	all	Service mode: ui, api, mcp, all
GPU_AUTO_SELECT	true	Auto-select GPU with lowest memory
API_HOST	0.0.0.0	API bind address
API_PORT	8000	API port
MCP_HOST	0.0.0.0	MCP bind address
MCP_PORT	8001	MCP port

🔌 API Endpoints

Endpoint	Method	Description
/health	GET	Health check
/api/gpu/status	GET	GPU status
/api/gpu/select	POST	Select GPU
/api/algorithms	GET	List algorithms
/api/backtest	POST	Run backtest
/api/jobs	GET	List jobs
/api/results/{id}	GET	Get results
/ws	WS	WebSocket for real-time updates

🤖 MCP Tools

Tool	Description
gpu_status	Get GPU information
gpu_select	Select specific GPU
list_algorithms	List available algorithms
run_backtest	Execute backtest
get_results	Retrieve backtest results

📁 Project Structure

docker/
├── Dockerfile           # Multi-stage build
├── docker-compose.yml   # Compose configuration
├── nginx.conf           # Reverse proxy config
├── supervisord.conf     # Process manager
├── app/
│   ├── api_server.py    # FastAPI REST API
│   ├── mcp_server.py    # MCP Server
│   ├── gpu_manager.py   # GPU auto-selection
│   └── static/
│       └── index.html   # Web UI
└── .env.example         # Environment template

🛠️ Tech Stack

• Base: QuantConnect Lean Engine
• Runtime: Python 3.10+, .NET 6
• API: FastAPI, Uvicorn
• GPU: NVIDIA CUDA, pynvml
• Frontend: Tailwind CSS, Alpine.js
• Process: Supervisord, Nginx

📝 License

Apache 2.0 - See LICENSE

⭐ Star History

Star History Chart

📱 Follow Us