~ $ inspect projects/maverick-mcp

Maverick MCP Stock Market Server

Open source stock analysis that runs on your computer

Maverick MCP is my open-source stock market server for the Model Context Protocol (MCP), the standard that lets an AI assistant call outside tools. It connects an assistant to Yahoo Finance data for stock analysis, screening, portfolio tracking, and Python backtesting. The server runs on your computer and keeps your portfolio, watchlists, and trade journal in a local database. Its core tools don't need an API key.

Updated .

cat project.json

At a glance

  • Creator and maintainer: I created it and maintain it as wshobson on GitHub
  • Tools: 38 core tools in release v1.2.1, plus 12 optional backtesting tools and 3 optional research tools
  • Market data: Yahoo Finance through yfinance, with no API key for core tools
  • Storage: local SQLite by default, with optional PostgreSQL and Redis
  • Clients: Claude Desktop, Claude Code, Codex, Cursor, VS Code, and other MCP clients
  • Language: Python 3.12 or later
  • License: MIT

ls tools/

What it does

AreaWhat you can do
Market data and technical analysisGet quotes, historical prices, fundamentals, RSI, MACD, and observed support and resistance levels
Stock screeningRun bullish, bearish, and supply and demand screens over the symbols you have loaded
Portfolio trackingRecord positions, calculate average cost and profit or loss, review risk, manage watchlists, and keep a trade journal
Optional backtesting and researchTest strategies on historical data, compare results, or research companies and sectors with source citations

The backtesting and research tools are optional packages. Research also needs a language model and a web search service that you configure, and those providers can charge for use.

cat DISCLAIMER

Limits

Maverick MCP is for education and information. It does not place orders or give financial advice, and a portfolio entry or journal trade is only a local record. Market data can be delayed, incomplete, or wrong, and backtests and technical indicators do not predict future returns.

run install

Run the server

Install Python 3.12 or later and uv. An MCP client that uses STDIO can then start the tagged release directly:

uvx --from "git+https://github.com/wshobson/maverick-mcp@v1.2.1" maverick-mcp --transport stdio

The GitHub repository covers installing from a clone, Docker, Streamable HTTP, the optional packages, and the configuration for each client.

ls related/