NautilusTrader MCP Server

MCP Server integration for the NautilusTrader algorithmic trading platform.

Overview

This project provides a Model Context Protocol (MCP) server that allows AI agents and other MCP clients to interact with NautilusTrader for:

• Live trading operations (order management, account info, positions)
• Backtesting trading strategies
• Accessing market data

Installation

1. Clone the repository:

   git clone <repository-url>
   cd nautilus-mcp

2. Create a virtual environment:

   python -m venv .venv
   source .venv/bin/activate  # On Windows use `.venv\Scripts\activate`

3. Install dependencies:

   pip install -r requirements.txt
   # Or for development:
   # pip install -e .[dev]  <- Assuming a setup.py with dev extras is added later

Usage

1. Configure the server: (Details TBD - likely via MCP settings or a config file)
2. Run the server:

   python -m nautilus_mcp

3. Register with MCP CLI (Optional but recommended):

   mcp install src/nautilus_mcp/server.py --name nautilus-trader -v # Add env vars if needed

Project Structure

nautilus-mcp/
├── .venv/                  # Virtual environment
├── docs/                   # Documentation files
├── src/
│   └── nautilus_mcp/
│       ├── __init__.py
│       ├── __main__.py     # Entry point for `python -m nautilus_mcp`
│       ├── server.py       # Core MCP Server class
│       ├── tools/          # MCP tools implementations
│       │   ├── __init__.py
│       │   └── (e.g., trading.py, backtest.py)
│       └── resources/      # MCP resource implementations
│           ├── __init__.py
│           └── (e.g., market_data.py)
├── tests/                  # Unit and integration tests
│   ├── __init__.py
│   └── (e.g., test_server.py)
├── .gitignore
├── README.md
└── requirements.txt

Development

(Add details about running tests, linting, contributing guidelines later)

License

(Specify license)