Polymarket MCP Server

A Model Context Protocol (MCP) compliant server that exposes the public Polymarket Central Limit Order Book (CLOB) REST API as tools that any LLM-powered IDE or agent can invoke. Compatible with Cursor, Continue (VS Code), GitHub Copilot Chat, JetBrains AI Assistant, OpenAI ChatGPT (Assistants API), Smithery and any other MCP-aware interface.

✨ Features

• 🪄 Universal MCP Support – works with Cursor, Continue, Copilot, JetBrains AI, Smithery and more
• ⚡ Python implementation – using native MCP protocol with async support
• 📈 Essential endpoints out-of-the-box – list markets, fetch single market details and more
• 🔌 Stateless & no auth required – queries Polymarket's public REST endpoints; no keys needed
• 🐳 Docker-ready – ship & deploy anywhere in seconds
• 📝 MIT licensed – free for personal and commercial use

🚀 Quick Start

Local Development

# Clone the repository
git clone https://github.com/CarlosIbCu/polymarket-mcp.git
cd polymarket-mcp

# Install dependencies
pip install -r requirements.txt

# Run the server
python -m polymarket_mcp.main

Testing with MCP Inspector

You can test the server using the official MCP Inspector:

# Install and run the MCP Inspector
npx @modelcontextprotocol/inspector python -m polymarket_mcp.main

This will:

1. Start the MCP server
2. Launch the inspector web interface
3. Provide a URL to access the inspector (usually http://localhost:6274)

The inspector allows you to:

• View available MCP tools
• Test tool calls interactively
• Inspect server capabilities
• Debug MCP protocol messages

Docker Deployment

# Build the image
docker build -t polymarket-mcp .

# Run the container
docker run -p 3333:3333 polymarket-mcp

Smithery Deployment

This server is ready for deployment on Smithery, the MCP hosting platform:

1. Fork/clone this repository
2. Connect your GitHub account to Smithery
3. Select this repository and choose "Custom Deploy"
4. Smithery will automatically detect the configuration and deploy

The smithery.yaml configuration includes:

• Container runtime settings
• Server configuration options
• Health checks
• Tool schemas

🛠 Available Tools

list_markets

List available prediction markets with optional limit parameter.

Parameters:

• limit (integer, optional): Maximum number of markets to return (default: 10)

Example:

{
  "limit": 5
}

get_market

Fetch detailed information about a specific market.

Parameters:

• market_id (string, required): Market ID to fetch

Example:

{
  "market_id": "0x9deb0baac40648821f96f01339229a422e2f5c877de55dc4dbf981f95a1e709c"
}

🔧 Configuration

Server configuration is handled via environment variables or Smithery config:

server:
  host: "0.0.0.0"  # Server host
  port: 3333       # Server port
  timeout: 30      # Request timeout in seconds

📚 API Documentation

• MCP Inspector: Use npx @modelcontextprotocol/inspector for interactive testing
• Polymarket API Docs: Official Documentation

🧪 Testing

# Test the Polymarket API connectivity
python test_server.py

# Run with MCP Inspector for interactive testing
npx @modelcontextprotocol/inspector python -m polymarket_mcp.main

🤝 Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

📄 License

This project is MIT licensed. See the LICENSE file for details.