Unofficial GuruFocus MCP Server & Python API Library

NOTE: This repo is currently a work in progress and is not yet ready for production use.

A comprehensive GuruFocus MCP server and Python API client library with Pydantic models for all responses.

Packages

This monorepo contains two packages:

gurufocus-api

A standalone Python client library for the GuruFocus API.

• Full API coverage with type hints
• Three-tier earnings-aware caching
• Rate limiting and retry logic
• Pydantic models for all responses
• Async-first design with httpx

gurufocus-mcp

An MCP server exposing GuruFocus data to AI assistants.

• FastMCP framework with HTTP/SSE transport
• 50+ tools for stocks, gurus, insiders, politicians, economic data, and more
• Resources for direct data access
• TOON format support for 30-60% token reduction in LLM contexts

Quick Start

Installation

# Install both packages
pip install gurufocus-api gurufocus-mcp

Using the API Client

from gurufocus_api import GuruFocusClient

async with GuruFocusClient() as client:
    summary = await client.get_summary("AAPL")
    print(summary.gf_score)

Running the MCP Server

export GURUFOCUS_API_TOKEN=your-token
gurufocus-mcp

Claude Desktop Integration

First, install the package:

pipx install gurufocus-mcp   # recommended for CLI tools
# or
pip install gurufocus-mcp

Then add to your Claude Desktop configuration file:

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

{
  "mcpServers": {
    "gurufocus": {
      "command": "gurufocus-mcp",
      "env": {
        "GURUFOCUS_API_TOKEN": "your-token-here"
      }
    }
  }
}

After adding the configuration, restart Claude Desktop. You should see the GuruFocus tools available in Claude's interface.

Running from source (for development):

{
  "mcpServers": {
    "gurufocus": {
      "command": "uv",
      "args": ["run", "gurufocus-mcp"],
      "env": {
        "GURUFOCUS_API_TOKEN": "your-token-here"
      }
    }
  }
}

Output Format (TOON)

The MCP server supports TOON format for token-efficient responses, achieving 30-60% reduction in token usage compared to JSON.

Per-request format selection:

All tools accept a format parameter:

• "toon" (default) - Token-efficient format optimized for LLM contexts
• "json" - Standard JSON format for debugging or compatibility

# TOON format (default)
result = await client.call_tool("get_stock_summary", {"symbol": "AAPL"})

# JSON format
result = await client.call_tool("get_stock_summary", {"symbol": "AAPL", "format": "json"})

Environment variable configuration:

# Change the default output format (default: toon)
export GURUFOCUS_DEFAULT_OUTPUT_FORMAT=json

Configuration

All settings can be configured via environment variables with the GURUFOCUS_ prefix.

Required

Variable	Description
GURUFOCUS_API_TOKEN	Your GuruFocus API token

Cache Settings

Variable	Default	Description
GURUFOCUS_CACHE_ENABLED	true	Enable response caching
GURUFOCUS_CACHE_DIR	.cache/gurufocus-mcp	Directory for cache storage

Recommended: Set an absolute path for GURUFOCUS_CACHE_DIR to ensure consistent caching across sessions:

export GURUFOCUS_CACHE_DIR="$HOME/.cache/gurufocus-mcp"

Or in Claude Desktop config:

{
  "mcpServers": {
    "gurufocus": {
      "command": "gurufocus-mcp",
      "env": {
        "GURUFOCUS_API_TOKEN": "your-token-here",
        "GURUFOCUS_CACHE_DIR": "/Users/yourname/.cache/gurufocus-mcp"
      }
    }
  }
}

Rate Limiting

Variable	Default	Description
GURUFOCUS_RATE_LIMIT_ENABLED	true	Enable rate limiting
GURUFOCUS_RATE_LIMIT_RPM	30.0	Maximum requests per minute
GURUFOCUS_RATE_LIMIT_DAILY	0	Maximum requests per day (0 = unlimited)

API Settings

Variable	Default	Description
GURUFOCUS_API_BASE_URL	https://api.gurufocus.com/public/user	API base URL
GURUFOCUS_API_TIMEOUT	30.0	Request timeout in seconds
GURUFOCUS_API_MAX_RETRIES	3	Maximum retry attempts for failed requests

Output Format

Variable	Default	Description
GURUFOCUS_DEFAULT_OUTPUT_FORMAT	toon	Default format: toon (token-efficient) or json

Logging

Variable	Default	Description
GURUFOCUS_LOG_LEVEL	INFO	Logging level: DEBUG, INFO, WARNING, ERROR
GURUFOCUS_LOG_FORMAT	console	Log format: console (human-readable) or json (structured)

Development

Setup

# Clone the repository
git clone https://github.com/yourorg/gurufocus-mcp.git
cd gurufocus-mcp

# Install uv if not already installed
curl -LsSf https://astral.sh/uv/install.sh | sh

# Install dependencies
uv sync --all-packages

# Run tests
uv run pytest

# Run tests with coverage
uv run pytest --cov --cov-report=html --junitxml=junit.xml

# Run linting
uv run ruff check .

Project Structure

gurufocus-mcp/
├── packages/
│   ├── gurufocus-api/     # Python API client
│   └── gurufocus-mcp/     # MCP server
├── docs/                  # Documentation
├── examples/              # Usage examples
└── pyproject.toml         # Workspace config

Disclaimer

This project is an unofficial tool and is not affiliated with, endorsed by, or sponsored by GuruFocus.com, LLC or any of its subsidiaries.

• Non-Affiliation: This software is developed independently and does not represent the views or opinions of GuruFocus.
• Trademarks: "GuruFocus" and the GuruFocus logo are trademarks of GuruFocus.com, LLC.
• Data Usage: Users are responsible for ensuring their use of this tool complies with the GuruFocus Terms of Use and any applicable API usage agreements. This tool is provided "as is" and intended for educational and personal use.
• No Warranty: The authors of this software make no warranty as to the accuracy, completeness, or reliability of the data retrieved using this tool.
• No Investment Advice: The information provided by this tool is for informational and educational purposes only and does not constitute investment advice, financial advice, trading advice, or any other sort of advice. You should not treat any of the tool's content as such.
• External Methodologies: References to third-party investment strategies, books, authors, or personalities (e.g., Phil Town, Rule #1, Warren Buffett) are for educational purposes to demonstrate the application of published methodologies. This project is not affiliated with, endorsed by, or sponsored by any referenced individuals or entities. All trademarks belong to their respective owners.

License

MIT