Red5d/me-mcp

Me MCP

Personal Information MCP Server

A self-hosted MCP (Model Context Protocol) server allowing AI assistants to access personal information that you supply, and receive contact requests.

Overview

This MCP server provides AI assistants with structured access to your personal details like bio, skills, and contact methods, while also offering a secure way for others to send you messages through your configured webhook.

Features

• 📋 Personal Info API: Expose your professional details to AI assistants
• 📬 Contact Mechanism: Let people reach you without directly sharing your email
• 🔐 Privacy Control: Configure exactly what information you want to share
• 🤖 AI-Friendly Interface: Pre-built prompts and resources for modern AI systems

Installation

1. Clone this repository:

   git clone https://github.com/yourusername/me-mcp.git
   cd me-mcp
   

2. Create and configure your config.json file:

   cp config.json.example config.json
   nano config.json  # Edit with your information
   

Configuration

Edit config.json with your personal information and webhook URL. The "name" field is required, all others are optional and can be added/removed as needed:

{
    "webhook_url": "https://your-webhook-endpoint.com",
    "personal_info": {
        "name": "Your Name",
        "handle": "yourhandle",
        "contact_methods": {
            "github": "https://github.com/yourhandle",
            "matrix": "https://matrix.to/#/@yourhandle:example.com",
            "X": "https://x.com/yourhandle",
            "email": "you@example.com"
        },
        "occupation": "Your Job",
        "skills": ["Your", "Skills", "Here"],
        "interests": ["Your", "Interests", "Here"],
        "bio": "A short biography about yourself."
    }
}

Usage

Run the MCP server:

uv run mcp run -t sse me_mcp.py

Your MCP server will be available via SSE at http://0.0.0.0:8000 which you can then expose to remote LLMs using a reverse proxy or similar.

You can also use mcp-proxy which includes a few more configuration options to run the server.

MCP Tools

• about_me(): Returns your configured personal information
• contact(): Allows sending you messages via webhook

MCP Resources

• about://me: JSON resource containing your personal information

MCP Prompts

• about_me_prompt(): Asks the AI to describe your information
• send_message(): Guides the AI to help someone contact you

Security Considerations

• Set up a secure webhook endpoint that processes contact requests
• Only expose information you're comfortable sharing publicly
• Consider running behind a reverse proxy with rate limiting

License

MIT