Overview

This repository contains the Hunter MCP (Model Context Protocol) server — a small adapter that lets any MCP-enabled LLM client (for example, Claude for Desktop) call Hunter API endpoints via natural-language prompts. The MCP server exposes Hunter API functionality as “tools” that an LLM can invoke, translating conversational intents into authenticated Hunter API requests and returning structured results back to the model.

Important: this project is deprecated and no longer maintained. All functionality has been migrated to Hunter’s Remote MCP Server. New users should start with the hosted Remote MCP option documented on Hunter’s API docs: https://hunter.io/api-documentation#mcp. The information below explains how the original local MCP server worked to help developers migrate or run a local instance for testing.

Features

Exposes Hunter B2B data (people and companies) to MCP-capable LLMs using conversational prompts
Wraps Hunter endpoints as discrete tools (Domain Search, Email Finder, Email Verifier, Enrichment, Lead Creation)
Simple configuration via an MCP server definition with environment variables for the API key
Designed for local use with MCP clients (e.g., Claude Desktop) or for integration into custom MCP setups

Installation / Configuration

Prerequisites:

Hunter API key (create in your Hunter dashboard)
Python 3.13+
The uv runner (https://github.com/astral-sh/uv)
MCP CLI/runtime installed for your LLM client (e.g., Claude Desktop MCP support)

Example MCP server config (JSON) for adding the local server to an MCP client:

{
  "mcpServers": {
    "hunter-mcp": {
      "command": "uv",
      "args": [
        "run",
        "--with",
        "mcp[cli]",
        "mcp",
        "run",
        "PATH_TO_hunter-mcp/main.py"
      ],
      "env": {
        "HUNTER_API_KEY": "YOUR_HUNTER_API_KEY"
      }
    }
  }
}

Install and run with the MCP tooling (example for Claude Desktop):

# from the hunter-mcp repo folder, substitute your API key
mcp install main.py -v HUNTER_API_KEY=YOUR_HUNTER_API_KEY
mcp run hunter-mcp

Environment variables:

HUNTER_API_KEY — your Hunter API key used for authenticated requests

If you prefer the maintained, hosted option, see Hunter’s Remote MCP documentation: https://hunter.io/api-documentation#mcp

Available Tools / Resources

The original MCP server exposed the following tools (each maps to a Hunter API endpoint or workflow):

Tool name	Purpose
Domain Search	Find email addresses associated with a domain
Email Finder	Locate a person’s email from name + domain
Email Verifier	Check deliverability and risk score for an email
Email Enrichment	Enrich an email/person with additional metadata
Company Enrichment	Fetch company details and firmographic data
Lead Creation	Save found contacts as Leads in Hunter workflows

Notes:

Not all Hunter API endpoints were wrapped by the original MCP server. The hosted Remote MCP server provides the most up-to-date coverage.
The GitHub repo: https://github.com/hunter-io/hunter-mcp contains the original source and license (MIT).

Use Cases

Investigative search via LLM:
- Prompt: “Find likely contact emails for the VP of Engineering at example.com and verify deliverability.”
- Flow: The model calls Domain Search + Email Finder, then Email Verifier. The MCP server sends requests to Hunter, aggregates results, and returns a concise summary (emails + verification status).
Enrich a contact during a conversation:
- Prompt: “I have [email protected] — give me any available firmographics, job title, and social links.”
- Flow: The model invokes Email Enrichment (and possibly Company Enrichment) to retrieve structured fields that the assistant can present to the user.
Create Leads from discovery:
- Prompt: “Add the verified contacts from acme.com to our Lead list with tags ‘outreach’ and ‘Q2’.”
- Flow: After discovery and verification, the model uses Lead Creation to persist those contacts using Hunter’s lead endpoints.

Example LLM-style prompt that triggers tool use:

Find email addresses for "Jane Doe" at "acme.com" and verify which emails are deliverable. If a deliverable email is found, create a Lead named "Jane Doe (Acme)".

Expected simplified output (what the MCP server returns to the model) might include:

found_emails: [“[email protected]”]
verification: { “[email protected]”: “deliverable” }
lead_creation: { “id”: “lead_12345”, “status”: “created” }

Migration & Maintenance

This repository is no longer updated. Developers should migrate to Hunter’s Remote MCP Server for production or long-term use.
For testing or to understand the integration pattern, you can inspect the original source at https://github.com/hunter-io/hunter-mcp and compare it to the Remote MCP documentation at https://hunter.io/api-documentation#mcp.

License

The original project is MIT-licensed. See the repo’s LICENSE file for details: https://github.com/hunter-io/hunter-mcp/blob/main/LICENSE.md

Hunter MCP Server: Functionality Moved to Remote