People MCP Server

A Python-based MCP (Model Context Protocol) server that enables AI assistants to search and retrieve people information from a Supabase database using natural language queries.

Features

• Natural Language Search: Search for people using natural language queries
• Detailed Person Information: Get comprehensive details about specific individuals
• Image Search: Find images associated with people by email
• List All People: Browse all people in the database with pagination

Tools Available

1. search_people

Search for people using natural language queries.

Examples:

• "find John Smith"
• "who works at Tech Corp"
• "person with email john@example.com"
• "people from San Francisco"
• "software engineer position"

2. get_person_details

Get comprehensive information about a specific person by their email address.

3. search_by_image

Find images associated with a person by their email address.

4. list_all_people

List all people in the database with an optional limit (default: 10, max: 100).

Installation

1. Install dependencies:

pip install -r requirements.txt

2. Run the server:

python src/server.py

The server will start on http://localhost:8000/mcp

Testing with MCP Inspector

1. Install MCP Inspector:

npm install -g @modelcontextprotocol/inspector

2. Run MCP Inspector:

npx @modelcontextprotocol/inspector

3. Connect to the server:

• Go to http://localhost:3000
• Select "Streamable HTTP" as transport
• Enter URL: http://localhost:8000/mcp
• Click "Connect"

Configuration

All configuration is hardcoded in src/server.py:

• Supabase URL: https://afwpcbmhvjwrnolhrocz.supabase.co
• Supabase Service Key: Hardcoded service role key with full access
• Tables Used:
  • nyne_results: Main people data table
  • images_names: Image storage table

Database Schema

nyne_results Table

• email (text) - Primary identifier
• name (text) - Full name
• phone (text) - Phone number
• company (text) - Company name
• title (text) - Job title
• location (text) - Geographic location
• data (json) - Extended structured data containing:
  • displayname, bio, gender
  • fullphone (array), altemails (array)
  • social_profiles (object)
  • organizations (array)

images_names Table

• email (text) - Person's email
• image (bytea) - Binary image data

Natural Language Query Processing

The server intelligently parses natural language queries to extract:

• Email addresses: Detected using regex patterns
• Phone numbers: Recognized in various formats
• Company names: Keywords like "works at", "from company"
• Job titles: Keywords like "title", "position", "role"
• Locations: Keywords like "from", "in", "location"
• Names: Default extraction when no specific fields detected

Deployment to Render

Option 1: One-Click Deploy

Click the "Deploy to Render" button above.

Option 2: Manual Deployment

1. Fork or push this repository to GitHub
2. Connect your GitHub account to Render
3. Create a new Web Service on Render
4. Connect your repository
5. Render will automatically detect the render.yaml configuration

Your server will be available at: https://your-service-name.onrender.com/mcp

Local Development

Setup Environment

conda create -n people-mcp python=3.13
conda activate people-mcp
pip install -r requirements.txt

Run Server

python src/server.py

Run with Custom Port

PORT=3000 python src/server.py

Example Usage

Once connected via MCP Inspector or an AI assistant:

1. Search for people:

   • Tool: search_people
   • Input: "find people who work at Google"

2. Get detailed information:

   • Tool: get_person_details
   • Input: "john.doe@example.com"

3. List people:

   • Tool: list_all_people
   • Input: 20 (to see 20 people)

Poke Integration

You can connect this MCP server to Poke at poke.com/settings/connections.

To test the connection explicitly, ask Poke something like: Tell the subagent to use the "People MCP" integration's "search_people" tool to find John Smith

License

MIT