Connecting 130M+ B2B Contacts to Claude Using MCP
The article discusses a Model Context Protocol (MCP) server that allows users to search and enrich 130M+ B2B contacts directly within AI assistants like Claude, Cursor, and Windsurf, without the need for external API calls or dashboards.
Why it matters
This MCP server provides a convenient way for AI assistants to access and leverage a large database of B2B contacts, enabling more powerful prospecting, research, and analysis capabilities.
Key Points
- 1MCP is an open standard for connecting AI assistants to external data sources, similar to USB-C for AI tools
- 2The MCP server provides two main tools: people_search (130M+ contacts with 20+ filters) and contact_enrich (get full profile details)
- 3The server is built using Supabase Edge Functions, mcp-lite, and Hono, with a simple architecture and around 200 lines of TypeScript
Details
The article explains how the author has built an MCP server that allows users to search and enrich a database of over 130 million B2B contacts directly within their AI assistants, such as Claude, Cursor, and Windsurf. The server exposes two main tools: people_search, which lets users search the contact database with over 20 different filters (job titles, company details, location, skills, etc.), and contact_enrich, which retrieves the full profile details for a specific person, including verified work and personal emails, phone numbers, social media links, and more. The server is built using Supabase Edge Functions, the mcp-lite library, and the Hono framework, with a simple architecture that requires only around 200 lines of TypeScript code. The author provides several real-world use cases for the server, such as prospecting, pre-meeting research, list building, and competitive analysis. The service is currently available for free for 12 months, with 500 credits per month and no credit card required.
No comments yet
Be the first to comment