A Model Context Protocol (MCP) server that exposes Binance cryptocurrency exchange data to Large Language Model (LLM) agents. This server allows LLMs to access real-time and historical market data from Binance through a standardized interface.
Features
Live Price Data: Get current prices for any trading pair on Binance
Binance Cryptocurrency Market Data MCP Service
2026-11-16
ดาว
—
เครื่องมือ
—
อัปไทม์
—
Order Book Access: Retrieve order book snapshots showing buy/sell interest at different price levels
Historical Price Data: Fetch OHLCV (Open, High, Low, Close, Volume) candlestick data for any timeframe
Real-time WebSocket Streams: Subscribe to real-time trade, ticker, and order book updates via WebSockets
Comprehensive Market Data: Access trades, 24hr statistics, aggregate trades, rolling window data, and more
Exchange Information: Access trading rules, symbol information, and fee structures
Read-only Operation: All data is fetched via Binance's public REST API (no API keys required)
MCP Standard Compliant: Works with any MCP-compatible LLM client
Requirements
Python 3.8+
mcp package with CLI tools (mcp[cli])
requests library for REST API
websockets library for WebSocket streams
uvicorn for serving (optional)
Installation
Clone this repository:
git clone https://github.com/yourusername/binance_mcp_server.git
cd binance_mcp_server
Install dependencies:
pip install -r requirements.txt
Usage
Running the Server Directly
To run the server in standalone mode:
python run_server.py
This will start the MCP server, which will listen for connections via STDIO.
Development Mode with MCP Inspector
For development and testing, use the MCP Inspector:
mcp dev run_server.py
This opens the MCP Inspector interface where you can test the server's tools interactively.
Installing in Claude Desktop
To use this server with Claude Desktop:
Install the MCP CLI tools if you haven't already:
pip install "mcp[cli]"
Install the server in Claude Desktop:
mcp install run_server.py
You can now access Binance data directly within your Claude Desktop conversations.
Example Client
An example client script is provided to demonstrate programmatic usage:
python example_client.py
This script connects to the server and retrieves various types of market data.
Available Tools
Connectivity and Basic Info
ping_binance(): Test connectivity to the Binance API server
Example: ping_binance()
get_server_time(): Get the current server time from Binance
Example: get_server_time()
Market Data
get_price(symbol): Get the current price for a trading pair
Example: get_price(symbol="BTCUSDT")
get_order_book(symbol, depth=10): Get the current order book
cleanup_all_streams(): Close all WebSocket connections and clean up resources
Example: cleanup_all_streams()
Project Structure
binance_mcp_server/
├── binance_api.py # Core REST API interaction module
├── binance_ws_api.py # WebSocket connection management
├── commands/ # MCP command definitions
│ ├── __init__.py
│ ├── market_data.py # Price/order book/historical data commands
│ ├── market_info.py # Exchange info and metadata commands
│ └── websocket_streams.py # WebSocket stream commands
└── server.py # Main MCP server setup and initialization
Examples
Basic Connectivity
# Check if Binance API is reachable
is_connected = ping_binance()
# Get the current server time (milliseconds since epoch)
server_time = get_server_time()
Getting Current Market Data
# Get the current price of Bitcoin
btc_price = get_price(symbol="BTCUSDT")
# Get detailed 24-hour statistics
btc_stats = get_24hr_ticker(symbol="BTCUSDT")
print(f"BTC price change: {btc_stats['priceChangePercent']}%")
print(f"BTC 24h volume: {btc_stats['volume']} BTC")
# Get rolling window statistics (4-hour window)
btc_4h_stats = get_rolling_window_ticker(symbol="BTCUSDT", window_size="4h")
print(f"BTC 4h price change: {btc_4h_stats['priceChangePercent']}%")
Working with WebSocket Streams
# Subscribe to real-time trade updates
trade_sub = subscribe_to_trade_stream(symbol="BTCUSDT")
# After some time, get the latest trade data
latest_trade = get_latest_stream_data(stream_name="btcusdt@trade")
print(f"Latest trade price: {latest_trade['data']['p']}")
# Subscribe to candlestick updates for chart data
kline_sub = subscribe_to_kline_stream(symbol="ETHUSDT", interval="5m")
# Clean up when done
unsubscribe_from_stream(stream_name="btcusdt@trade")
unsubscribe_from_stream(stream_name="ethusdt@kline_5m")
Implementation Status
For a detailed overview of the implemented and planned API endpoints, refer to the API Implementation Status document.
Extending the Server
Adding New Tools
To add new tools, create appropriate functions in the Binance API module and then register them as MCP tools in one of the command modules.
Supporting Private API Endpoints
To support authenticated API calls:
Modify binance_api.py to include authentication
Create additional command modules for trading operations
Register the new commands in server.py
Troubleshooting
Connection Issues: Ensure the server is running before attempting to connect with a client
Rate Limiting: Binance may rate-limit excessive API calls; consider implementing caching for high-traffic deployments
WebSocket Stability: WebSocket connections may disconnect after 24 hours (Binance limit); the server will attempt to reconnect automatically
Data Format: Different symbols or intervals may return data in slightly different formats
License
This project is licensed under the MIT License - see the LICENSE file for details.
Acknowledgments
Binance for providing a comprehensive public API
The MCP project for standardizing LLM tool interactions