A production-ready Model Context Protocol server for cryptocurrency market intelligence, powered by real-time exchange data, machine learning, and multi-agent orchestration.
- 📊 Real-Time Exchange Data - Binance, Kraken, Coinbase orderbooks & tickers
- 🔬 Market Microstructure - OFI, OBI, Microprice, VPIN analytics
- 🤖 ML Price Prediction - DeepLOB-Lite model for buy/sell signals
- 🎯 Trading Strategies - Multi-signal aggregation engine
- 👥 Multi-Agent System - Research, Risk, Execution agents with voting
- 📡 WebSocket Streaming - Real-time orderbook/ticker updates
- 🔔 Smart Alerts - Price-based notifications with background monitoring
- 💼 Portfolio Management - Risk analysis, P&L tracking, paper trading
- 🕵️ Anomaly Detection - Spoofing, layering, market regime classification
- 📈 Interactive Dashboard - Streamlit UI for live market visualization
- 🌐 Sentiment Analysis - Fear & Greed Index integration
# Clone repository
git clone https://github.com/Arshad-13/CryptoIntel-MCP.git
cd CryptoIntel-MCP
# Create virtual environment
python -m venv venv
source venv/bin/activate # On Windows: venv\Scripts\activate
# Install dependencies
pip install -r requirements.txt-
Copy configuration to Claude Desktop:
cp docs/claude_desktop_config.json %APPDATA%\Claude\claude_desktop_config.json
-
Update paths in the config file to match your installation
-
Restart Claude Desktop
-
Test:
"Fetch orderbook for BTC/USDT" "Run analysis pipeline for ETH/USDT with sentiment 0.7"
streamlit run dashboard.pyYou: "What's the current liquidity situation for BTC/USDT?"
Claude: [Fetches orderbook, calculates depth, analyzes spread]
"The BTC/USDT orderbook shows strong liquidity with
$2.3M in bids within 0.5% of mid-price..."
from tools.strategy_tools import get_trading_signal
signal = await get_trading_signal('ETH/USDT', sentiment_score=0.6)
# Returns: {'signal': 'BUY', 'confidence': 0.82, ...}You: "Run full analysis on SOL/USDT"
Claude: [Orchestrates Research → Risk → Execution agents]
"Research Agent: ML prediction BUY (78% confidence)
Risk Agent: Position size approved (2x BTC)
Execution Agent: Recommended entry: $142.35"
- Dashboard: Live orderbook depth charts, ML predictions, portfolio P&L
- WebSocket Streams: Subscribe to orderbook/ticker updates
- Alerts: Get notified when BTC crosses $90,000
┌─────────────────────────────────────────────────────────┐
│ Claude Desktop │
└────────────────────┬────────────────────────────────────┘
│ JSON-RPC / STDIO
┌────────────────────▼────────────────────────────────────┐
│ Market Intelligence MCP Server │
│ │
│ ┌──────────────┐ ┌──────────────┐ ┌──────────────┐ │
│ │ Exchange │ │ Analytics │ │ Strategy │ │
│ │ Tools │ │ Engine │ │ Engine │ │
│ └──────┬───────┘ └──────┬───────┘ └──────┬───────┘ │
│ │ Direct HTTP │ ML Models │ Agents│ │
│ ┌──────▼──────────────────▼──────────────────▼───────┐ │
│ │ Binance │ Kraken │ Coinbase │ WebSockets │ │
│ └──────────────────────────────────────────────────────┘ │
└─────────────────────────────────────────────────────────┘
Key Technologies:
- 🐍 Python 3.13 - Async I/O, type hints
- 🔌 httpx - Direct REST API calls (no CCXT overhead)
- ⚡ websockets - Real-time streaming
- 🧠 ONNX Runtime - ML model inference
- 📊 Streamlit - Interactive dashboard
- 💾 SQLite - Local persistence
See Architecture Documentation for details
| Document | Description |
|---|---|
| Installation Guide | Setup, configuration, troubleshooting |
| API Reference | Complete tool documentation with examples |
| Architecture | System design, data flows, scalability |
| Dashboard Guide | Dashboard features and customization |
| Changelog | Version history and feature timeline |
CryptoIntel-MCP/
├── core/ # Business logic (analytics, ML, risk)
├── tools/ # MCP tool implementations
├── agents/ # Multi-agent system
├── tests/ # Test suite (pytest)
├── docs/ # Documentation
├── dashboard.py # Streamlit UI
└── market_server.py # MCP server entry point
pytest tests/ -vCoverage: 13 test files, 100+ test cases
- Create function in
tools/your_tool.py - Register in
market_server.py:@mcp.tool() def your_tool(param: str) -> str: return your_function(param)
- Add tests in
tests/test_your_tool.py
# Optional: For premium APIs
CRYPTO_API_KEY=your_coingecko_api_keyAutomatic failover: Binance → Kraken → Coinbase
Configure in tools/exchange_tools.py:
EXCHANGE_FALLBACK_ORDER = ["binance", "kraken", "coinbase"]{
"symbol": "BTC/USDT",
"exchange": "binance",
"bids": [[88360.79, 0.5], [88360.0, 1.2]],
"asks": [[88361.0, 0.3], [88361.5, 0.8]],
"fallback_used": false
}{
"signal": "BUY",
"confidence": 0.85,
"components": {
"ml_prediction": "buy",
"ml_confidence": 0.78,
"sentiment_score": 0.7,
"risk_reward_ratio": 3.2
}
}{
"final_decision": "BUY",
"confidence": 0.82,
"agents": {
"research": {"recommendation": "buy", "confidence": 0.78},
"risk": {"approved": true, "max_size": 0.05},
"execution": {"entry_price": 88360.0, "slippage": 0.02}
}
}Contributions welcome! Please:
- Fork the repository
- Create a feature branch (
git checkout -b feature/AmazingFeature) - Commit changes (
git commit -m 'Add AmazingFeature') - Push to branch (
git push origin feature/AmazingFeature) - Open a Pull Request
This project is licensed under the MIT License - see the LICENSE file for details.
- Model Context Protocol (MCP) - Anthropic's extensible AI integration framework
- DeepLOB - Limit order book prediction research
- Alternative.me - Fear & Greed Index data
- Issues: GitHub Issues
- Discussions: GitHub Discussions
- Documentation: docs/
Built with ❤️ for the crypto trading community
Disclaimer: This is an educational project. Not financial advice. Trade at your own risk.