22 checks the page check judges from outside, plus one only the agent on the site can answer — in the order an engine meets them: can it reach the pages, can it read them, does it know who you are, do others vouch for you. Each page says what is checked, why, how to see it on your own site and how to fix it — and whether the agent does it for you.
robots.txt, CDN/bot filter, login walls, noindex, HTTP status, sitemap, canonical
robots.txt allows AI crawlers
Check 0.2No CDN or bot filter blocks AI crawlers
Check 0.3No login or paywall blocks the content
Check 0.4 · agent writes itAI crawler visits are being seen
Check 0.5No page is marked noindex
Check 0.6 · agent switches itsitemap.xml is present and valid
Check 0.7All pages return HTTP 200
Check 1.3Canonical tags point to the page itself
text without JavaScript, response time, internal links, answer-first structure, self-contained sections, headings/dates/authorship
Text is present without JavaScript
Check 1.2Response time is within limits
Check 1.4Pages link to each other
Check 3.1 · agent writes itPages open with the answer
Check 3.2 · agent writes itSections are self-contained
Check 3.3 · agent writes itHeadings, dates and authorship are present
JSON-LD on the page, entity links (@id, sameAs), Wikidata record, consistent external profiles
JSON-LD identifies the organization
Check 2.2 · agent switches itStructured data is in the raw HTML, not injected by scripts
Check 2.3 · agent switches itStructured data entities are linked (@id)
Check 2.4 · agent switches itsameAs links to external profiles are present
Check 2.5 · agent writes itStructured data matches the visible text
Check 4.1Brand has a Wikidata record and a Wikipedia article
Check 4.2External profiles match the brand
comparison pages, press and reviews about your category that AI engines read
Free, from outside, in minutes — with the exact finding behind every line.