Skip to main content

How to connect Hunter to Claude with MCP

Connect Hunter to Claude using the Model Context Protocol (MCP) so you can search for emails, find leads, and run domain searches without leaving your chat.

Hunter integrates with Claude through the Model Context Protocol (MCP), letting you search for email addresses, find leads, and run domain searches using natural language, without leaving your Claude session.

This article walks you through connecting Hunter to Claude, whether you're using the Claude web or desktop app.

If you came here from the quick-start on hunter.io/mcp, that page shows the three-step version:

If you were logged in, Step 1 showed your own API key with a copy button, instead of the link in this screenshot.

Step 2 there — "Add the remote MCP server" — uses the same server URL as Method 1 below. The difference: Claude's connector setup logs you in with one-click OAuth, so you won't need to paste an API key there. You'll only need the API key from Step 1 if you're using Method 2 (the config file).

Prerequisites

To connect Hunter to Claude, you'll need:

  • A Hunter account: create one for free if you don't have one

  • An API key from your Hunter dashboard: only needed if you're using Method 2 below; Method 1 authenticates with one-click OAuth instead

Add the server

There are a few ways to add Hunter as an MCP server in Claude, depending on your plan and role.

Quickest: start from your Hunter account

If you're logged in to Hunter, you can skip most of the manual setup:

  1. Go to "Integrations" in your Hunter account, then the "AI integrations" tab.

  2. Make sure "Claude" is selected under "Pick your AI tool".

  3. Click "Add to Claude".

That opens Claude with the Hunter connector details already filled in. You click "Add", then "Connect", and log in to Hunter.

It's the same connector as Method 1 below, so the same plan and role rules apply. If you'd rather set it up by hand, or you need the API key route, both are below.

Method 1: Claude connector settings

This is the easiest way to connect Hunter to Claude. It works in both the Claude web app and the Claude desktop app.

💡 On Claude Team and Enterprise plans, only workspace Owners can add custom connectors, and they add it from "Organization settings" then "Connectors". Once an Owner has added Hunter, everyone else in the workspace just clicks "Connect" — there's nothing else to set up.

If you're not an Owner and nobody can add it for you, use Method 2 instead. That route needs the Claude desktop app, so it isn't an option if you only use Claude in a browser.

  1. Open Claude.

  2. Click "Customize" in the left sidebar.

  3. Select "Connectors".

  4. Click "Add custom connector".

  5. Fill in the following:

    • Name: Hunter

    • Remote MCP server URL: https://mcp.hunter.io/mcp

  6. Click "Add".

  7. Click "Connect".

  8. Log in and authorize access in the Hunter login window that opens.

Once connected, start a new chat. You can ask Claude things like: "Use Hunter to find the email pattern for stripe.com" and it will respond using your Hunter account.

Method 2: Configuration file (Claude Desktop)

Use this method if you're a non-Owner member on a Claude Team or Enterprise plan, or if you prefer to authenticate with your API key directly.

  1. Open Claude Desktop.

  2. Go to "Settings" → "Developer" → "Edit Config" to open your claude_desktop_config.json file.

  3. Add the following inside mcpServers:

    { "mcpServers": { "hunter": { "command": "npx", "args": [ "-y", "mcp-remote", "https://mcp.hunter.io/mcp", "--header", "X-API-Key:${HUNTER_API_KEY}" ], "env": { "HUNTER_API_KEY": "YOUR_HUNTER_API_KEY" } } } }

  4. Replace YOUR_HUNTER_API_KEY with your key from your Hunter dashboard.

  5. Save the file.

  6. Restart Claude Desktop fully (quit and reopen, don't just close the window).

👉 This method keeps your API key out of your command line and process list by passing it through an environment variable (env), rather than inline in the request header. It requires Node.js to be installed on your machine, since it uses npx to run mcp-remote.

Set your tool permissions

Connecting Hunter adds a set of Hunter tools to Claude, and each one has a permission level. "Always allow" lets Claude use that tool without asking. "Needs approval" means Claude asks you first, every time.

What we recommend:

  • Read-only tools (looking things up): Always allow

  • Write and delete tools (saving, updating or deleting leads): Needs approval

That lets Claude run your prospecting without stopping at every search, while you keep the final say on anything that changes your Hunter account. Verifying an email address is the step that uses credits, and how many depends on your plan — the current figure is in your Hunter account.

To set them, go to "Customize" then "Connectors" then "Hunter" and set each tool.

How to revoke Hunter's access

Connecting Hunter to Claude gives it ongoing access to your Hunter account. You can cut that access from Hunter at any time:

  1. Go to "Integrations" in your Hunter account.

  2. Open the "Authorized apps" tab.

  3. Find Claude in the list and click "Revoke".

The list shows when each app was connected and when it was last used. If you've connected the same app more than once, it appears once per connection — revoke each entry you want to cut off. You can connect again later whenever you need to.

This is separate from removing the Hunter connector inside Claude: revoking here acts on your Hunter account directly.

How to confirm the connection is working

Start a new chat and ask: "Use Hunter to find the email pattern for stripe.com."

If Hunter returns results, you're all set.

Troubleshooting

If Hunter isn't responding to Claude after connecting, check the following:

  • Make sure you started a new chat after connecting — MCP connections aren't picked up mid-conversation.

  • If you used Method 2, confirm you fully restarted Claude Desktop (quit and reopen, don't just close the window).

  • Double-check that your API key is correct and active in your Hunter dashboard.

Building your own integration

If you're wiring Hunter into your own agent rather than using the connector above, Hunter publishes reference files written for an assistant to read directly:

  • hunter.io/llms.txt — the index. A short list of every agent-facing file, product page and API reference.

  • hunter.io/llms-full.txt — the full reference. Every tool on Hunter's hosted MCP server, how authentication works, and how those tools map onto the REST API.

  • hunter.io/claude.md — the steps on this page, written for Claude to walk you through. Paste the link into a chat and Claude takes you through the setup.

You don't need any of these to use the methods above.

Still stuck? Our support team can help you get set up.

Contact our support team
Related Articles
Did this answer your question?