MonteWalk 🚀

Institutional-grade quantitative finance tools for AI agents.

MonteWalk is an MCP server that gives AI assistants (like Claude) access to professional trading tools: real-time market data, paper trading, risk analysis, and portfolio optimization.

---

✨ Key Features

• 📊 Market Intelligence: Aggregates data from Yahoo Finance, Alpaca, and CoinGecko.
• 💰 Paper Trading: Real execution via Alpaca ($100k virtual account).
• ⚠️ Risk Management: Monte Carlo simulations, VaR, and volatility metrics.
• 📈 Backtesting: Test strategies with walk-forward analysis.
• 🗞️ News & Sentiment: Multi-source news aggregation with NLP sentiment scoring.
• 🤖 Intelligent Watchlist: Auto-syncs with your portfolio via AI prompts.

---

🚀 Quick Start

1. Prerequisites

• Python 3.12+
• Alpaca Account (Free Paper Trading)
• NewsAPI Key (Optional)

2. Installation

git clone https://github.com/N-lia/MonteWalk.git
cd MonteWalk
python3 -m venv .venv
source .venv/bin/activate
pip install mcp yfinance pandas numpy scipy pandas_ta textblob gnews newsapi-python pycoingecko alpaca-py
python -m textblob.download_corpora

3. Configuration

Create a .env file:

cp .env.example .env

Edit .env with your keys:

ALPACA_API_KEY=your_key
ALPACA_SECRET_KEY=your_secret
NEWSAPI_KEY=your_key  # Optional

4. Run Server

For Hugging Face Deployment (Gradio UI + MCP):

uv run app.py

For Local Claude Desktop (stdio-based MCP):

uv run server.py

---

🔌 Connect to Claude Desktop

Local Connection (Recommended):

Add to your claude_desktop_config.json:

{
  "mcpServers": {
    "montewalk": {
      "command": "/absolute/path/to/MonteWalk/.venv/bin/python",
      "args": ["/absolute/path/to/MonteWalk/server.py"]
    }
  }
}

Note: Replace /absolute/path/to/MonteWalk with your actual path.

Alternative (Streamable HTTP via Gradio): If using app.py, configure with URL:

{
  "mcpServers": {
    "montewalk": {
      "url": "http://localhost:7860/gradio_api/mcp/"
    }
  }
}

⚠️ Note: You must start app.py BEFORE launching Claude Desktop for this to work.

---

📚 Usage

Just ask Claude:

• "What's my portfolio risk?"
• "Analyze TSLA technicals and sentiment."
• "Backtest a 10/50 MA crossover on Apple."
• "Buy 10 shares of MSFT."
• "Run morning briefing." (Auto-syncs watchlist!)

See API_REFERENCE.md for full tool documentation.

---

🛠️ Architecture

• Core: FastMCP server
• Data: yfinance, Alpaca, CoinGecko, NewsAPI
• Analysis: NumPy, SciPy, Pandas, TextBlob
• Security: Paper trading only, local data storage

---

📄 License

MIT License. Built for the MCP 1st Birthday Hackathon.