geo-tool-check

Is this page readable for AI search? Checks crawler access, structure and citability for ChatGPT, Perplexity, Claude and Google AI — as a CLI and as an MCP server.

Everything runs on your machine. No account, no API key, no data sent to geo-tool.com.

CLI

npx geo-tool-check example.com

https://example.com/
92/100  Excellent  534 words · 238 ms

  directAnswers      25/25  ██████████████████
  structure          15/20  ██████████████░░░░
  schemaMarkup       15/15  ██████████████████
  citations          12/15  ██████████████░░░░
  multimedia          5/5   ██████████████████
  platformSpecific   20/20  ██████████████████

In CI

npx geo-tool-check https://example.com/pricing --min-score 70

Exit code 1 below the threshold. Catches the deploy that blocks GPTBot or hides the content behind JavaScript.

npx geo-tool-check example.com --crawlers   # only robots.txt, exits 1 if a scored crawler is blocked
npx geo-tool-check example.com --json       # machine-readable
npx geo-tool-check example.com --lang de    # German labels

MCP server

Gives an agent three tools for AI visibility.

{
  "mcpServers": {
    "geo-tool-check": {
      "command": "npx",
      "args": ["-y", "geo-tool-check", "--mcp"]
    }
  }
}

Or point directly at the binary: npx -y geo-tool-check-mcp.

Also on Smithery:

npx -y smithery mcp add tracktronaut/geo-tool-check

To rebuild and republish the bundle:

npm run bundle                    # geo-tool-check.mcpb
./scripts/publish-smithery.sh     # needs: smithery auth login

| Tool | What it does | Network | |---|---|---| | check_ai_readiness | Full page score, six categories, findings | fetches the page | | check_ai_crawlers | Which AI crawlers robots.txt blocks | fetches robots.txt | | check_citability | Scores a passage while drafting | none |

What the score means

Twenty-five checks across six categories, 0 to 100. The same code that powers geo-tool.com — the scoring logic is imported, not reimplemented, so both produce the same number for the same input.

Node executes no JavaScript. This tool therefore reports what a plain HTML crawler sees — which is exactly what GPTBot, ClaudeBot and PerplexityBot see. A client-rendered page will score low here, and that is the finding, not a limitation. To compare the crawler's view against the rendered page, use the browser extension.

There is deliberately no JavaScript rendering fallback: a paid rendering service is exactly the running cost this package must not create.

Zero dependencies

The MCP protocol over stdio is JSON-RPC with three methods, implemented directly. Nothing to install beyond Node 18+, no supply chain, no version drift.

License

MIT. Built by track by track GmbH.