hackmyip-mcp

The no-key IP intelligence MCP — paste one URL, no signup.

https://hackmyip.com/mcp

That's the whole setup. No API key. No account. No OAuth dance. No credit card. Point any MCP client at that URL and your agent can immediately look up IPs, geolocate addresses, classify VPN/datacenter traffic, query DNS, run WHOIS, check spam blacklists, audit security headers, scan ports and test email authentication.

It's a thin, stateless wrapper around the free HackMyIP public REST API (docs) running on Cloudflare Workers at the edge. Every tool passes the API's JSON straight back to your agent — nothing is cached, rewritten, or invented.

Install

Claude Code

claude mcp add --transport http hackmyip https://hackmyip.com/mcp

Claude Desktop

Settings → Connectors → Add custom connector, then paste:

https://hackmyip.com/mcp

{
  "mcpServers": {
    "hackmyip": {
      "command": "npx",
      "args": ["-y", "mcp-remote", "https://hackmyip.com/mcp"]
    }
  }
}

Cursor

~/.cursor/mcp.json (global) or .cursor/mcp.json (per project):

{
  "mcpServers": {
    "hackmyip": {
      "url": "https://hackmyip.com/mcp"
    }
  }
}

Codex CLI

~/.codex/config.toml:

[mcp_servers.hackmyip]
url = "https://hackmyip.com/mcp"

Restart Codex. No codex mcp login step — this server is authless.

VS Code / GitHub Copilot

.vscode/mcp.json:

{
  "servers": {
    "hackmyip": {
      "type": "http",
      "url": "https://hackmyip.com/mcp"
    }
  }
}

Anything else

Any client that speaks MCP Streamable HTTP works. Endpoint: https://hackmyip.com/mcp (mirror: https://hackmyip-mcp.shitcoinape.workers.dev/mcp). Both the 2025-06-18 and 2026-07-28 protocol revisions are served.

Tools

Things worth knowing

• my_ip reports the IP the MCP request arrived from. If your client runs on your laptop, that's your real public IP — which is exactly what makes "is my VPN actually working?" answerable. If it runs on a remote host or behind a corporate proxy, you get that egress IP instead. The tool says so in its response.
• ip_blacklist_check never fakes a clean result. Lists that refuse public queries come back as unavailable, so read total_checked next to listed_count.
• port_scan is a real scan from Cloudflare's network. Only point it at hosts you're authorized to test. Private and reserved ranges are rejected upstream.
• whois_domain depends on third-party RDAP registries which occasionally 5xx. A failure there is upstream, not a malformed query — retry.

Fair use

The upstream API allows 60 requests/minute per client IP (20/min for port_scan and security_headers_check), resetting every 60 seconds. There is no paid tier to upgrade to and no key that lifts it — it exists so the service stays free for everyone. If you hit it, tools return the upstream 429 message and you should back off.

Prefer bulk_ip_lookup over a loop of ip_lookup calls.

Privacy

This server stores nothing. It holds no session state, writes no logs of your queries, and requires no identity. Each tool call is a stateless pass-through to https://hackmyip.com/api/*.

Run it yourself

git clone https://github.com/0xvibly/hackmyip-mcp
cd hackmyip-mcp
npm install
npm run dev          # local dev server at http://localhost:8787/mcp
npm run deploy       # deploy to your own Cloudflare account

Built with @modelcontextprotocol/server and Cloudflare's agents createMcpHandler — stateless, no Durable Objects, no database.

Links

• Website — https://hackmyip.com
• API docs — https://hackmyip.com/api-docs
• 50+ browser tools (VPN leak tests, fingerprinting, breach checks) — https://hackmyip.com/tools

License

MIT