Polymarket MCP
A Model Context Protocol (MCP) server for Polymarket prediction markets, providing real-time market data, prices, and AI-powered analysis tools for Claude Desktop integration.
README
Polymarket MCP Server
A Model Context Protocol (MCP) server for Polymarket prediction market data, built with Rust and the official MCP Rust SDK. This server provides real-time market data, prices, and betting information from Polymarket through MCP tools, resources, and prompts.
Features
- Real-time Market Data: Fetch active markets, trending markets, and market details
- Market Search: Search markets by keywords across questions, descriptions, and categories
- Price Information: Get current yes/no prices for any market
- MCP Resources: Auto-refreshing market data resources
- MCP Prompts: AI-powered market analysis and arbitrage detection
- Caching: Built-in caching with configurable TTL for performance
- Configuration: Flexible configuration via environment variables or TOML files
Prerequisites
- Rust 1.70+ (install via rustup)
- Polymarket API credentials (optional for basic read-only access)
Installation
For Claude Desktop
-
Clone the repository:
git clone https://github.com/0x79de/polymarket-mcp cd polymarket-mcp -
Build the release binary:
cargo build --release -
Configure Claude Desktop:
Edit your Claude Desktop configuration file at:
- macOS:
~/Library/Application Support/Claude/claude_desktop_config.json - Windows:
%APPDATA%\Claude\claude_desktop_config.json - Linux:
~/.config/Claude/claude_desktop_config.json
Add the Polymarket MCP server:
{ "mcpServers": { "polymarket": { "command": "/full/path/to/polymarket-mcp/target/release/polymarket-mcp", "env": { "RUST_LOG": "info" } } } } - macOS:
-
Restart Claude Desktop to load the new MCP server.
Optional Configuration
Set up environment variables for enhanced functionality:
cp .env.example .env
# Edit .env with your API credentials (optional for basic usage)
Configuration
Environment Variables
Copy .env.example to .env and configure as needed:
# API Configuration
POLYMARKET_API_KEY=your_api_key_here
POLYMARKET_API_BASE_URL=https://gamma-api.polymarket.com
# Trading API Credentials (for advanced operations)
POLYMARKET_SECRET=your_secret_here
POLYMARKET_PASSPHRASE=your_passphrase_here
# Wallet Credentials (for blockchain interactions)
WALLET_PUBLIC_KEY=0x_your_public_key_here
WALLET_PRIVATE_KEY=your_private_key_here
# Server Settings
POLYMARKET_CACHE_ENABLED=true
POLYMARKET_CACHE_TTL=60
POLYMARKET_LOG_LEVEL=info
Configuration File
Alternatively, create a config.toml file:
[server]
name = "Polymarket MCP Server"
max_connections = 100
timeout_seconds = 30
[api]
base_url = "https://gamma-api.polymarket.com"
api_key = "your_api_key_here"
timeout_seconds = 30
max_retries = 3
[cache]
enabled = true
ttl_seconds = 60
max_entries = 1000
[logging]
level = "info"
format = "pretty"
enable_colors = true
Usage
Claude Desktop Integration
To use with Claude Desktop, configure your claude_desktop_config.json:
{
"mcpServers": {
"polymarket": {
"command": "/path/to/polymarket-mcp/target/release/polymarket-mcp",
"env": {
"RUST_LOG": "info"
}
}
}
}
Configuration file location:
- macOS:
~/Library/Application Support/Claude/claude_desktop_config.json - Windows:
%APPDATA%\Claude\claude_desktop_config.json - Linux:
~/.config/Claude/claude_desktop_config.json
Building for Production
# Build release version for Claude Desktop
cargo build --release
# The binary will be at: target/release/polymarket-mcp
Standalone Server (Development)
# Run with default configuration (development mode)
cargo run
# Run with debug logging
RUST_LOG=debug cargo run
# Test the server manually (not recommended for production)
echo '{"jsonrpc":"2.0","method":"initialize","params":{"protocolVersion":"2024-11-05","capabilities":{},"clientInfo":{"name":"test","version":"1.0.0"}},"id":0}' | cargo run
MCP Tools
The server provides the following MCP tools:
get_active_markets
Fetch currently active prediction markets.
Parameters:
limit(optional): Maximum number of markets to return (default: 50)
Example:
{
"tool": "get_active_markets",
"arguments": {
"limit": 10
}
}
get_market_details
Get detailed information about a specific market.
Parameters:
market_id(required): The ID of the market
Example:
{
"tool": "get_market_details",
"arguments": {
"market_id": "0x123..."
}
}
search_markets
Search markets by keyword.
Parameters:
keyword(required): Search termlimit(optional): Maximum results (default: 20)
Example:
{
"tool": "search_markets",
"arguments": {
"keyword": "election",
"limit": 5
}
}
get_market_prices
Get current yes/no prices for a market.
Parameters:
market_id(required): The ID of the market
get_trending_markets
Get markets with highest trading volume.
Parameters:
limit(optional): Maximum number of markets (default: 10)
MCP Resources
Auto-refreshing data resources:
markets:active- List of active marketsmarkets:trending- Trending markets by volumemarket:{id}- Specific market details
MCP Prompts
AI-powered analysis prompts:
analyze_market
Comprehensive market analysis including sentiment, liquidity, and trading opportunities.
Arguments:
market_id(required): Market to analyze
find_arbitrage
Detect arbitrage opportunities across related markets.
Arguments:
keyword(required): Search term for related marketslimit(optional): Number of markets to analyze
market_summary
Overview of top prediction markets with trading recommendations.
Arguments:
category(optional): Filter by categorylimit(optional): Number of markets to include
API Documentation
Market Object Structure
{
"id": "0x123...",
"slug": "market-slug",
"question": "Will X happen by Y date?",
"description": "Detailed market description",
"active": true,
"closed": false,
"liquidity": 50000.0,
"volume": 125000.0,
"end_date": "2024-12-31T23:59:59Z",
"outcomes": ["Yes", "No"],
"outcome_prices": ["0.65", "0.35"],
"category": "Politics"
}
Error Handling
The server implements comprehensive error handling:
- Network Errors: Automatic retry with exponential backoff
- Rate Limiting: Built-in rate limiting with configurable limits
- Data Validation: All API responses are validated
- Caching: Prevents redundant API calls
Development
Running Tests
# Run all tests
cargo test
# Run tests with output
cargo test -- --nocapture
# Run specific test
cargo test test_market_client
Code Structure
src/
├── main.rs # MCP server implementation
├── config.rs # Configuration management
├── models.rs # Data structures and types
└── polymarket_client.rs # Polymarket API client
Adding New Features
- New Tools: Add methods to
PolymarketMcpServerinmain.rs - New Models: Define structures in
models.rs - API Endpoints: Extend
PolymarketClientinpolymarket_client.rs - Configuration: Add options to
Configinconfig.rs
Deployment
Docker
# Build Docker image
docker build -t polymarket-mcp .
# Run container
docker run -d \
--name polymarket-mcp \
-e POLYMARKET_API_KEY=your_key \
-p 8080:8080 \
polymarket-mcp
Systemd Service
Create /etc/systemd/system/polymarket-mcp.service:
[Unit]
Description=Polymarket MCP Server
After=network.target
[Service]
Type=simple
User=polymarket
ExecStart=/usr/local/bin/polymarket-mcp
Restart=always
RestartSec=10
Environment=RUST_LOG=info
EnvironmentFile=/etc/polymarket-mcp/.env
[Install]
WantedBy=multi-user.target
Troubleshooting
Common Issues
JSON Parsing Errors in Claude Desktop Logs
Error: Unexpected token '', " [2m2025-0"... is not valid JSON
Solution: This indicates logs are being written to stdout instead of stderr. The server has been configured to write all logs to stderr to avoid interfering with MCP JSON communication.
MCP Validation Errors
Error: Invalid input, Expected string, received null in id field
Solution: The server now properly handles JSON-RPC notifications (like notifications/initialized) by not responding to them, which is the correct behavior.
Server Connection Issues
Error: Server disconnected or Server transport closed unexpectedly
Troubleshooting steps:
- Verify the binary path in your
claude_desktop_config.json - Check that the binary exists:
ls -la /path/to/target/release/polymarket-mcp - Test the binary manually:
echo '{"jsonrpc":"2.0","method":"initialize","params":{"protocolVersion":"2024-11-05","capabilities":{},"clientInfo":{"name":"test","version":"1.0.0"}},"id":0}' | /path/to/target/release/polymarket-mcp - Check logs in Claude Desktop:
~/Library/Logs/Claude/mcp.log(macOS)
Build Issues
Error: cannot find function 'stdin' in module 'tokio::io'
Solution: Ensure tokio has the io-std feature enabled in Cargo.toml:
tokio = { version = "1.0", features = ["rt-multi-thread", "macros", "sync", "time", "signal", "io-std"] }
Debug Mode
Enable debug logging to see detailed MCP communication:
# In your claude_desktop_config.json
{
"mcpServers": {
"polymarket": {
"command": "/path/to/polymarket-mcp/target/release/polymarket-mcp",
"env": {
"RUST_LOG": "debug"
}
}
}
}
Log Locations
- Claude Desktop MCP Logs:
~/Library/Logs/Claude/mcp.log(macOS) - Claude Desktop Main Logs:
~/Library/Logs/Claude/main.log(macOS) - Server stderr output: Appears in Claude Desktop MCP logs
Monitoring
The server provides structured logging and can be monitored using:
- Logs: JSON or pretty-printed logs with configurable levels (written to stderr)
- Metrics: Built-in caching metrics and API call tracking
- Health: Connection pooling and retry statistics
- MCP Protocol: Full request/response logging in debug mode
Contributing
- Fork the repository
- Create a feature branch:
git checkout -b feature-name - Make your changes and add tests
- Run tests:
cargo test - Run lints:
cargo clippy - Format code:
cargo fmt - Submit a pull request
Security
- API keys are handled securely and never logged
- Private keys are optional and only used for trading operations
- All HTTP connections use TLS
- Input validation prevents injection attacks
License
This project is licensed under the MIT License - see the LICENSE file for details.
Support
- Issues: GitHub Issues
- Documentation: MCP Specification
- Polymarket API: Official Documentation
Changelog
v0.1.1 (Current)
- Fixed: JSON parsing errors in Claude Desktop by redirecting logs to stderr
- Fixed: MCP validation errors with notification handling
- Fixed: Added
io-stdfeature to tokio for stdin/stdout support - Improved: Proper JSON-RPC protocol compliance
- Added: Comprehensive troubleshooting documentation
- Added: Claude Desktop integration guide
v0.1.0
- Initial release with core MCP functionality
- Support for market data fetching and searching
- MCP resources and prompts implementation
- Comprehensive configuration system
- Built-in caching and error handling
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。