Data Detective

Case file #001 · Generative Engine Optimization

What do AI crawlers actually see on your page?

Your page looks fine in the browser. But GPTBot, ClaudeBot and PerplexityBot don't use your browser. Most of them don't run JavaScript, some get stopped by your bot protection, and some were locked out by a robots.txt nobody remembers writing. Data Detective finds out — for people and for their agents.

Free page audit

Basic SEO, LLM readability and structured data, measured on the raw HTML — the version AI crawlers get. Plus the site's AI policy: llms.txt, robots.txt per AI crawler, Content Signals and page directives.

10 free audits per minute. We fetch the URL and the site's robots.txt and llms.txt once and keep nothing but the request in our logs.

The clues

The free audit tells you what's on the page. The clues tell you who can read it.

free

GEO Page Audit

Title, description, headings, robots, canonical, language, JSON-LD with required fields, author and date, Open Graph, JavaScript weight. Score out of 100.

GET /api/audit?url=

free

llms.txt & AI policy

Is there a valid llms.txt, which AI crawlers does robots.txt let in, what do the Content Signals allow, and does the page opt out with noindex or noai?

GET /api/check/ai-policy?url=

soon

Citation check

Does an LLM name your brand for the questions your customers ask?

soon

Case report

All findings as a readable report with concrete schema and content fixes.

Built for agents

Every clue is a plain HTTP call with JSON in and JSON out. Without a key, a paid clue answers 402 Payment Required with its price — the same handshake that pay-per-call protocols like x402 build on.

curl -X POST https://datadetective.app/api/clue/bot-view \
  -H "Authorization: Bearer $DD_KEY" \
  -d '{"url":"https://example.com/","rendered":{"textLength":4200}}'

The product list is machine-readable at /api/catalog, and there's an llms.txt for the agents reading this page. API keys are in private beta while payments are being built: request one by email.

In your browser: GEO Page Audit

The free GEO Page Audit extension scores the page you're on — Basic SEO, LLM readability, structured data — on the rendered page, locally in your browser. Then “Call the detective” runs the Bot-View clue and compares what you see with what the crawlers get. Only the URL and a text length are sent, and only when you click.

Chrome and other Chromium browsers · Firefox 140+ on desktop · Firefox 142+ on Android. The audit is free; the Bot-View clue needs an API key.

GEO Page Audit popup scoring a page out of 100 across Basic SEO, LLM readability and structured data.
One click scores the current page out of 100.
The Bot-View clue in the extension: GPTBot disallowed by robots.txt, ClaudeBot blocked with HTTP 403 by bot protection.
Bot-View: robots.txt, server response and visible text per AI crawler.

Questions

Why does raw HTML matter for AI search?

Most AI crawlers don't run JavaScript. Content that only appears after rendering is invisible to them and can't be cited in AI answers.

Is a blocked test request proof that a crawler is blocked?

No. We send each crawler's published user agent. Sites that verify bots by IP may block our test but let the real crawler through. A block is a strong hint, not proof, and every result says so.

Who is behind this?

Data Detective is a product from LLMO Consulting. It is the very first product on the leanplugins.dev store core delivering plugins to supercharge your AI agents.