Transparency by design

How AI Crawl Score Works

Scan stages, evidence boundaries and limitations of public website diagnostics.

A safe public request

We validate the URL, preflight DNS, check robots.txt, retrieve HTML and discover a sitemap. Requests identify as AICrawlScoreBot. We do not impersonate a verified crawler.

A bounded sample

The initial budget is five pages, fourteen retrieval attempts including redirects, one megabyte per response and a forty-second scan deadline. Requests are sequential. Sitemap traversal is deliberately shallow.

Explainable observations

Reports include findings, recommendations, source links, confidence and weights. Read test coverage together with the score. Unknown evidence is not a pass.

Limits of the method

Raw HTML cannot establish computed visibility, keyboard behavior, real crawler visits, indexing or citations. Browser comparisons and official search integrations are separate capabilities. A generic fetch does not prove access from any particular crawler infrastructure.


Scoring methodology · Data sources · Free tools