Crypto MCP Trading Server

A Model Context Protocol (MCP) server for LIVE cryptocurrency trading via Binance.US, enabling AI-driven trading with real-time market data, comprehensive analytics, crypto news, and intelligent position management.

Features

Core Trading

• ✅ MCP Protocol: Standard stdio-based MCP communication
• 🔐 Binance.US Integration: Direct API connection for live trading
• 💹 Advanced Order Types: Market, Limit, Stop-Loss, OCO (One-Cancels-Other)
• 🎯 Automatic TP/SL: Set take-profit and stop-loss in one order
• 💰 Smart Position Management: Easy position closing with OCO cancellation
• 💾 Local Database: SQLite tracks all trades with full OCO linking

Market Intelligence

• 📊 Comprehensive Market Data: Price, volume, liquidity metrics, volatility analysis
• 📈 Technical Analysis: RSI, MACD, SMA, EMA, Bollinger Bands with buy/sell signals
• 🕯️ Advanced Candlestick Analysis: Trend detection, support/resistance, pattern recognition
• 📰 Crypto News: CryptoPanic integration for market sentiment (24h delay)
• 📉 Order Book Depth: Real-time bid/ask analysis
• 🔍 Symbol Details: Trading rules, lot sizes, price filters

Analysis & Intelligence

• 🧠 AI-Powered Recommendations: Comprehensive position analysis
• 📊 Multi-Timeframe Analysis: 1m, 5m, 15m, 30m, 1h, 4h, 1d, 1w
• 🎯 Technical Sentiment: Aggregated buy/sell signals from all indicators
• 📈 Trend Detection: Identify UPTREND, DOWNTREND, SIDEWAYS markets
• 💧 Liquidity Scoring: Assess market depth and spread quality
• 🚀 Live Trading Only: All data from real Binance API (no paper mode)

Architecture

┌─────────────┐         ┌──────────────────┐         ┌─────────────┐
│             │  MCP    │                  │  REST   │             │
│  LLM Client │◄───────►│   MCP Server     │◄───────►│ Binance.US  │
│ (Claude/LM  │ stdio   │   (Python)       │   API   │   Exchange  │
│   Studio)   │Protocol │                  │         │             │
└─────────────┘         └──────────────────┘         └─────────────┘
                               │                  │
                               │                  │ News
                               │                  ▼
                               │           ┌──────────────┐
                               │           │ CryptoPanic  │
                               │           │     API      │
                               │           └──────────────┘
                               │
                               │ Logs All Trades
                               ▼
                        ┌─────────────┐
                        │   SQLite    │
                        │  Database   │
                        │ (trading.db)│
                        └─────────────┘

Installation

Prerequisites

• Python 3.8+
• Binance.US account with API keys
• pip package manager

Quick Install

pip install binance-mcp

Configuration

Add to your Claude Desktop or LM Studio configuration file:

macOS: ~/Library/Application Support/Claude/claude_desktop_config.json Windows: %APPDATA%\Claude\claude_desktop_config.json

{
  "mcpServers": {
    "crypto-trading": {
      "command": "binance-mcp",
      "env": {
        "BINANCE_API_KEY": "your_api_key_here",
        "BINANCE_API_SECRET": "your_api_secret_here",
        "CRYPTOPANIC_API_KEY": "optional_news_api_key"
      }
    }
  }
}

Note: Get a free CryptoPanic API key at https://cryptopanic.com/developers/api/

Available Tools

The AI has access to 14 powerful trading tools:

Market Intelligence

1. get_market_data - Comprehensive market intelligence

• Real-time price, volume, 24h stats
• Liquidity metrics (bid/ask depth, spread analysis)
• Volatility & volume trend analysis
• Technical indicators (RSI, MACD, SMA, EMA, Bollinger Bands)
• Automated buy/sell signals
• Trading activity metrics

2. get_candlestick_data - Advanced timeframe analysis

• OHLCV data for any timeframe (1m to 1w)
• Trend detection (UPTREND, DOWNTREND, SIDEWAYS)
• Support/resistance level calculation
• Candle pattern analysis (BULLISH, BEARISH, DOJI)
• Strength indicators (STRONG, MODERATE, WEAK)
• Timeframe-specific technical indicators

3. get_crypto_news - Market sentiment & trending coins

• Latest crypto news (24-hour delay)
• Sentiment analysis
• Trending cryptocurrencies
• CryptoPanic integration

4. get_all_markets - All tradable cryptocurrency pairs

5. get_order_book - Live order book depth analysis

6. get_24h_ticker - 24-hour statistics for symbols

7. get_symbol_details - Trading rules, lot sizes, filters

8. get_fee_structure - Trading fees for symbols

Account & Trading

9. get_account_balances - View your Binance account balances

10. place_order - Execute live trades

• Market, Limit, Stop-Loss orders
• Automatic OCO (One-Cancels-Other) for TP/SL
• Supports both LONG and SHORT positions
• Auto-formats quantity to exchange requirements

11. close_position - Smart position closing

• Automatically sells full or partial balance
• Cancels related OCO orders
• Market execution for immediate fills

12. get_open_orders - See all active orders

13. cancel_order - Cancel open orders

14. analyze_position - AI analyzes when to sell

• Current P&L calculation
• Technical indicator analysis
• Clear buy/hold/sell recommendations

Trading History

15. get_trade_history - Recent trade history

How It Works

1. Market Research with Intelligence

AI: "Analyze Bitcoin's market conditions"
→ Uses get_market_data("BTCUSDT")
→ Returns:
  💰 PRICE DATA:
    Current: $67,234.50 (+2.3%)
    24h Range: $65,100 - $68,500

  💧 LIQUIDITY:
    Liquidity Score: 87/100
    Spread: 0.02%

  🎯 TECHNICAL ANALYSIS:
    Overall Sentiment: BULLISH
    Buy Signals: 4 | Sell Signals: 1
    RSI: 58 [NEUTRAL]
    MACD: 120 [BUY]
    SMA_20 > SMA_50 [BUY]

2. Timeframe-Specific Analysis

AI: "Show me the 4-hour chart for Bitcoin"
→ Uses get_candlestick_data("BTCUSDT", timeframe="4h")
→ Returns:
  📈 TREND ANALYSIS:
    Trend: STRONG UPTREND
    Support: $66,100
    Resistance: $68,800

  🕯️ RECENT CANDLES:
    10/07 12:00 🟢 BULLISH (STRONG) | +1.8%
    10/07 08:00 🟢 BULLISH (MODERATE) | +0.5%
    10/07 04:00 🔴 BEARISH (WEAK) | -0.3%

3. Check Market News

AI: "What's the latest news on Bitcoin?"
→ Uses get_crypto_news(symbols=["BTC"])
→ Returns:
  📰 Crypto News (10 items):
  ⚠️ NOTE: News has 24-hour delay

  • Bitcoin ETF sees $200M inflow
    Source: CoinDesk | 🟢 Positive

  • Major institution adopts BTC
    Source: Bloomberg Crypto | 🟢 Positive

4. Place Trade with Auto TP/SL

AI: "Buy 0.01 BTC with take-profit at $70,000 and stop-loss at $65,000"
→ Uses place_order(
    symbol="BTCUSDT",
    side="BUY",
    order_type="MARKET",
    quantity=0.01,
    take_profit_price=70000,
    stop_loss_price=65000
  )
→ Returns:
  ✅ LIVE Order with OCO Exit Executed:
  ENTRY ORDER:
    Order ID: 12345
    BUY 0.01 BTCUSDT @ $67,234

  OCO EXIT ORDER:
    Take-Profit: $70,000
    Stop-Loss: $65,000
    Status: NEW

5. Close Position

AI: "Close my entire Bitcoin position"
→ Uses close_position(symbol="BTCUSDT")
→ Returns:
  ✅ Position Closed:
    Quantity Sold: 0.01 BTC
    Executed at: $68,500
    Status: FILLED
    OCO Order canceled.

Order Types & Strategies

Market Orders

• Execute immediately at current market price
• Best for quick entry/exit

Limit Orders

• Execute at specific price or better
• Good for patient entries

OCO Orders (One-Cancels-Other)

• Automatically place take-profit AND stop-loss
• When one executes, the other cancels
• Perfect for long and short positions

Example LONG with OCO:

# Buy BTC, auto-exit at profit OR loss
place_order(
  symbol="BTCUSDT",
  side="BUY",
  order_type="MARKET",
  quantity=0.01,
  take_profit_price=70000,  # Exit at profit
  stop_loss_price=65000      # Exit at loss
)

Example SHORT with OCO:

# Sell BTC short, auto-exit at profit OR loss
place_order(
  symbol="BTCUSDT",
  side="SELL",
  order_type="MARKET",
  quantity=0.01,
  take_profit_price=65000,   # Exit at profit (lower)
  stop_loss_price=70000       # Exit at loss (higher)
)

Technical Indicators

Supported Timeframes

• 1m, 5m, 15m, 30m, 1h, 4h, 1d, 1w

Available Indicators

RSI (Relative Strength Index)

• 14-period RSI
• < 30: Oversold (BUY signal)

• 70: Overbought (SELL signal)

MACD (Moving Average Convergence Divergence)

• 12/26/9 periods
• MACD > Signal: BUY
• MACD < Signal: SELL

SMA (Simple Moving Average)

• 20 and 50 periods
• SMA20 > SMA50: Golden Cross (BUY)
• SMA20 < SMA50: Death Cross (SELL)

EMA (Exponential Moving Average)

• 12 and 26 periods
• Used for MACD calculation

Bollinger Bands

• 20-period, 2 std dev
• Price < Lower Band: BUY
• Price > Upper Band: SELL

Database Tracking

All trades are automatically logged to trading.db (SQLite):

TradeLog Table

• Order ID, Symbol, Side, Type
• Quantity, Price, Fill Price
• Status, Timestamp
• OCO Tracking: Links entry/exit orders
• TP/SL Prices: Records target prices
• Strategy Notes: AI decision reasoning

DailyPnL Table

• Daily profit/loss tracking
• Win/loss ratio
• Total volume
• Trade count

Each computer maintains its own local trade history with full OCO order linking.

Position Analysis

The AI analyzes your positions and recommends when to sell based on:

Technical Indicators

• RSI - Overbought (>70) suggests selling
• MACD - Bearish crossover suggests selling
• Bollinger Bands - Price above upper band suggests selling
• Moving Averages - Death cross suggests selling

Profit/Loss Thresholds

• >10% profit - Consider taking profits
• >5% loss - Consider stop-loss
• Multiple sell signals - Strong recommendation to sell

Recommendations Include

• Clear action: SELL, HOLD, or TAKE PROFITS
• Suggested take-profit levels
• Suggested stop-loss levels
• Current P&L calculation
• Technical indicator breakdown

Security

• API keys stored in environment variables
• Never committed to version control
• MCP stdio protocol (no network exposure)
• Token-based authentication
• Request validation
• LOT_SIZE auto-formatting prevents order errors

Example Conversation

User: "Find me a good crypto to buy right now"

AI: Let me analyze the market for you...
[Uses get_crypto_news() to check sentiment]
📰 Latest News: Bitcoin ETF inflows, bullish sentiment

[Uses get_all_markets() to see available coins]
[Uses get_market_data() for top coins with full analysis]

Market Intelligence Analysis:

BTCUSDT:
💰 Price: $67,234 (+2.3%)
💧 Liquidity: 87/100 (Excellent)
📈 Volatility: 3.2% (Moderate)
🎯 Sentiment: BULLISH (4 buy, 1 sell)

ETHUSD:
💰 Price: $3,456 (+4.1%)
💧 Liquidity: 82/100 (Good)
📈 Volatility: 4.8% (High)
🎯 Sentiment: STRONG BULLISH (5 buy, 0 sell)

Based on comprehensive analysis, Ethereum shows the strongest signals with 5/5 buy indicators and strong momentum. Would you like me to buy ETH with automatic take-profit and stop-loss?

User: "Yes, buy $1000 worth with 10% take-profit and 5% stop-loss"

AI: [Uses get_account_balances() to verify funds]
✅ Balance confirmed: $5,000 available

[Calculates: $1000 / $3456 = 0.289 ETH]

[Uses place_order() with OCO]
✅ Order Executed:

ENTRY: Bought 0.289 ETH @ $3,456 = $1,000
OCO EXIT SET:
  Take-Profit: $3,802 (+10%)
  Stop-Loss: $3,283 (-5%)

I'll monitor this position. The OCO order will automatically:
- Sell at $3,802 for 10% profit, OR
- Sell at $3,283 to limit loss to 5%

Use `analyze_position` anytime to check status!

What's New in v3.0.0

Major Features

• ✨ Crypto News Integration - CryptoPanic API for market sentiment
• 📊 Enhanced Market Intelligence - Comprehensive liquidity & volatility metrics
• 🕯️ Advanced Candlestick Analysis - Trend detection, support/resistance, patterns
• 🎯 Smart Position Closing - close_position tool with OCO cancellation
• 📈 Consolidated Market Data - All indicators in one comprehensive tool

Improvements

• 🔧 LOT_SIZE Auto-Formatting - Prevents order rejection errors
• 💾 Enhanced Database - OCO tracking, TP/SL linking, strategy notes
• 🗑️ Removed Friction - Eliminated risk manager and paper trading mode

Breaking Changes

• ❌ Risk management removed (trade freely)
• ❌ Paper trading removed (live-only)
• ✅ All risk decisions now AI-driven

Development

Project Structure

Crypto MCP/
├── crypto_mcp_server/
│   ├── server.py              # MCP server & tool handlers
│   ├── binance_client.py      # Binance.US + news integration
│   ├── database.py            # SQLAlchemy models with OCO tracking
│   ├── config.py              # Configuration
│   └── schemas.py             # Pydantic models
├── pyproject.toml             # Package configuration
└── README.md                  # This file

Building from Source

# Clone repository
git clone https://github.com/yourusername/crypto-mcp.git
cd crypto-mcp

# Install in development mode
pip install -e .

# Build package
python -m build

# Upload to PyPI
python -m twine upload dist/*

Troubleshooting

"Binance.US client failed to initialize"

• Check API keys in configuration
• Ensure keys have trading permissions
• Verify API keys are for Binance.US (not Binance.com)

"Filter failure: LOT_SIZE"

• Fixed in v3.0.0! Quantities now auto-formatted
• Update to latest version: pip install --upgrade binance-mcp

"Connection closed" error

• Restart Claude Desktop or LM Studio
• Verify binance-mcp is installed: pip show binance-mcp
• Check configuration file path

AI doesn't see tools

• Restart client completely
• Verify MCP server is in config
• Check logs for server startup

News not showing

• Add CRYPTOPANIC_API_KEY to configuration
• Free key at https://cryptopanic.com/developers/api/
• Falls back to CoinGecko trending if no key

License

MIT License

Disclaimer

⚠️ LIVE TRADING WITH REAL MONEY

This software executes REAL trades on Binance.US with REAL money. Use at your own risk.

• Cryptocurrency trading is highly volatile
• You can lose your entire investment
• Past performance doesn't guarantee future results
• The authors are not responsible for financial losses
• News data has a 24-hour delay - not for time-sensitive decisions
• Always start with small amounts
• Never trade more than you can afford to lose

This is not financial advice. Trade responsibly.

Support

For issues or questions:

1. Check this README
2. Review trading.db for trade history
3. Check server logs for errors
4. Verify Binance.US API status
5. Visit: https://pypi.org/project/binance-mcp/

Links

• PyPI: https://pypi.org/project/binance-mcp/
• Binance.US API: https://docs.binance.us/
• CryptoPanic API: https://cryptopanic.com/developers/api/
• MCP Protocol: https://modelcontextprotocol.io/