&mcp Server Information

The &mcp server provides AI assistants with access to public &tag profiles through the Model Context Protocol (MCP). &tags are unique identifiers that connect people and companies with their publicly available information, preferences, and context.

What is an &tag?

An &tag is like a social media handle but for AI. Every person and company can have an &tag (like &dan or &nike) that contains structured information about them - preferences, interests, professional details, and context. This data is designed to be read by AI assistants to provide personalized responses.

Installations for AI Assistants

Many AI assistants (Claude, ChatGPT, Gemini, etc.) now support MCP server configuration. MCP servers are primarily used for developer endpoints, but adding a trusted MCP server to your AI assistant can give it some superpowers. Only paid AI assistant accounts (e.g., Claude Pro) can add MCP at this time. After adding the MCP server to your account, but sure to reload your app or browser client.

Claude (Recommended)

  1. Open https://claude.ai/settings/connectors
  2. Choose "Add custom connector"
  3. Name: &mcp
  4. Remote MCP server URL: https://mcp.and.com
  5. Ignore "Advanced settings"
     

ChatGPT

  • Coming soon


Gemini

  • Doesn't yet support MCP servers in their assistant

Using the MCP Server

Once installed, your AI assistant can:

  • Retrieve public &tag profiles: "Get information about &dan"
  • Search across &tags: "Find people interested in AI"
  • Access preferences and context: "What are &sarah's food preferences?"
  • Discover connections: "Show me &tags related to climate tech"

Developer Guide

Full documentation coming soon. 

Building with &Tag Data

Integrate public &tag data into your applications using the MCP protocol.

Quick Start:

  1. Install MCP SDK for your language (Python, TypeScript, Java, C#)
  2. Connect to: https://mcp.and.com
  3. Use available tools:
    • get_user_signals: Retrieve a specific &tag profile
    • search_user_signals: Search across &tags
    • list_all_users: Browse available &tags


Example (Python):

from mcp import ClientSession 
async with ClientSession("mcp.and.com") as session: 
	profile = await session.call_tool("get_user_signals", 					
		{"username": "dan"}
	)

Authentication:

Public &tag data requires no authentication. Private or premium features may require API keys in the future.

Rate Limits:

Standard rate limits apply:

  • 100 requests per hour for unauthenticated access
  • 1000 requests per hour for authenticated developers

Feelback and Support

Questions, issues, or feature requests: 
Feedback Form: and.com/feedback 
Documentation: coming soon
Status Page: coming soon

Technical Details

Protocol Version: MCP 2025-03-26 
Transport: HTTPS with SSE 
Response Format: JSON 
Available Tools: get_user_signals, search_user_signals, list_all_users, create_andeer_from_public_data 
Supported Clients: Claude (web, desktop, mobile, code), Gemini CLI, Custom MCP clients

Privacy and Data

All data served through this MCP server is public-facing information that &tag owners have explicitly made available. Users control their &tag data through and.com privacy settings.

Data includes: Professional information, preferences, interests, public social links Data excludes: Private messages, financial information, authentication credentials

Roadmap

Coming soon:

  • ChatGPT support
  • Gemini native client support
  • Real-time &tag updates via webhooks
  • Advanced search and filtering
  • Company and brand &tag enrichment
  • Developer analytics dashboard

 

Last updated: 10/24/25