headstarter

The Headstarter LinkedIn Network MCP Server exposes a collection of MCP tools that let AI assistants query and analyze LinkedIn profile data from the Headstarter network. Core capabilities include executing read-only SQL-like queries against the linkedIn network table, retrieving individual profiles by username or URN, and performing advanced searches with multiple filters. Specialized search tools extend this by enabling lookups by location, identifying profiles that are open to work, actively hiring, creators and thought leaders, and Headstarter-affiliated members. In addition, there are resources describing the database schema and statistics to help understand data shape and coverage. You can use these tools to assemble targeted lists of profiles, perform demographic or role-based analyses, and support recruitment, networking, or community-building efforts. Leverage the included examples to format your requests, such as listing Headstarter alumni in a specific city or locating hiring managers at particular companies. The server is designed to work with the MCP framework, enabling consistent tool invocation across AI agents and clients.

Prerequisites: Node.js installed (recommended LTS), npm or pnpm, a PostgreSQL database with the hs_linkedin_network schema populated from Headstarter data, and Redis for SSE transport. Steps: 1) Clone the repository for the Headstarter MCP server. 2) Create and configure environment variables (see env notes). 3) Install dependencies: npm install or yarn install. 4) Build and start the server (as per the project’s Start script, typically npm run build && npm run start). 5) Verify the server is running and accessible at the configured URL. 6) Register the MCP server with Cursor or MCP client tooling using the provided JSON snippet. 7) Optional: enable Fluid Compute in your hosting environment if supported by your deployment target.

Environment variables: DATABASE_URL (PostgreSQL DSN for hs_linkedin_network), REDIS_URL (Redis instance for SSE transport). Ensure your database and Redis are secured and accessible by the deployment. The MCP server enforces read-only access to queries and applies automatic query limits for safety. Use the provided example commands to test basic functionality before integrating into a larger workflow. If you encounter deployment issues on Vercel or other hosting platforms, confirm that the Next.js MCP adapter is properly configured and that the server route is accessible via the expected URL. For performance, enable Fluid Compute and adjust maxDuration if your hosting plan supports it.