Search 1 billion profiles from any AI agent with the People Search MCP Server

Connect Claude Desktop, Cursor, VS Code, or any MCP-compatible AI agent to the HeroHunt People Search API. Natural language search across GitHub, Stack Overflow, and the web with real-time data, AI screening, and contact enrichment. One config line, 10 tools, zero integration code.

Get your API key
Get started in minutes

Three steps to search talent from your AI agent

No SDK installation required for most clients. Add a single JSON config block and start searching.

Get your API key

Get your API key

Sign up at herohunt.ai and generate an API key from the dashboard. Free tier available for evaluation and small usage.

Add MCP config

Add the MCP config to your AI tool

Paste a single JSON block into your AI tool's MCP settings. Remote HTTP is recommended (no package install needed). For stdio-only clients, use npx herohunt-mcp.

Search with natural language

Search with natural language

Ask your AI agent to find candidates using plain English. The MCP server handles search, pagination, enrichment, and account management through 10 built-in tools.

1

Zero integration code, just JSON config

2

1B+ profiles, real-time data, AI screening

3

Works with Claude, Cursor, ChatGPT, and more

What you get

10 MCP tools covering search, enrichment, and account management

Natural language search
Natural language search

Describe the candidate in plain English or paste a job description. No boolean syntax required.

Contact enrichment
Contact enrichment

Add SMTP-verified email addresses and phone numbers to profiles on demand.

Dual transport
Dual transport

Connect via remote Streamable HTTP URL (zero install) or npx herohunt-mcp for stdio-only clients.

Credit-based pricing
Credit-based pricing

1 credit per reserved profile slot at search time. Pagination is free. Enrichment adds 1 credit per enriched profile. Free tier available.

All 10 MCP tools

Full toolset for talent search and account management

Check
people_search — natural language search across GitHub, Stack Overflow, and the web. Returns ranked, AI-screened profiles with structured data.
Check
people_search_paginate — fetch additional result pages for an existing search. Free unless enrichment is enabled.
Check
account_usage_get — returns current plan, total and remaining credits, rate limits, and next credit reset date.
Check
account_searches_list — paginated history of past searches with first-page profiles. Resume or review prior sourcing runs.
Check
account_upgrade_link — returns a dashboard URL for upgrading the plan directly from your AI agent.
Check
account_billing_portal_link — returns the Stripe billing portal URL for managing subscriptions and payment methods.
Check
api_keys_list — lists all API key prefixes, assigned scopes, and last-used timestamps.
Check
api_keys_create — creates a new API key with optional name, scopes, and expiry. Full key returned once only.
Check
api_keys_rotate — deactivates an existing key and issues a new one with identical scopes.
Check
api_keys_revoke — permanently revokes an API key. Irreversible.
FAQ

Frequently asked questions

MCP (Model Context Protocol) is an open standard that lets AI agents connect to external tools and data sources through a unified interface. The HeroHunt MCP Server exposes our People Search API as a set of tools that any MCP-compatible AI agent can call directly, without custom integration code.
Any MCP-compatible client works. This includes Claude Desktop, Cursor, ChatGPT Desktop, Windsurf, VS Code (via GitHub Copilot), Claude Code, and custom agents built with the MCP SDK. Remote HTTP transport works with most modern clients. For stdio-only clients, use npx herohunt-mcp.
For most AI tools, no. Claude Desktop, Cursor, and ChatGPT Desktop support remote HTTP endpoints natively. You just add a JSON config block pointing to https://api.herohunt.ai/api/v1/mcp with your API key. For stdio-only clients, run npx herohunt-mcp (requires Node.js 20+).
The MCP Server uses the same underlying People Search API and the same data. The difference is the interface: the REST API is for developers building applications with HTTP requests, while the MCP Server lets AI agents call search, enrichment, and account tools directly through the MCP protocol. Same data, same pricing, different access method.
You ask your AI agent something like "Use HeroHunt to find 5 senior Go engineers in Berlin and summarize the top 3." The agent calls the people_search tool with your natural language query, receives structured profile data (name, title, company, location, skills, platform links), and presents the results. To get contact info, ask "Enrich the results with emails and phone numbers."
The MCP Server uses the same credit-based pricing as the REST API. A new search charges 1 credit per reserved slot up front, where slots = min(maxResults, availableCredits) (default maxResults = 100, max 500). Paginating the same searchId is free: the maxResults reservation was already billed on the creation call. With enrich: true, +1 credit per LinkedIn profile on the returned page that comes back with an email or phone (a profile counts when it is returned with contact data). Non-LinkedIn profiles are never charged for enrichment. A free tier is available for evaluation and small usage. Paid plans offer flat-rate monthly pricing with higher volumes.
Yes. They share the same API key, the same credits, and the same account. Use the MCP Server for AI-agent workflows and the REST API for application integrations. Searches from either surface appear in the same account history.
Yes. You can connect programmatically using the MCP SDK. Import StreamableHTTPClientTransport from @modelcontextprotocol/sdk, point it to https://api.herohunt.ai/api/v1/mcp with your API key, and call tools like people_search directly from your Node.js agent. See the documentation in your API dashboard for a complete code example.