Raw HTML Availability

AI Readiness · Evidence-backed

What this check reads on your page

Words present in the HTML the server returns, before any JavaScript runs, are what this factor counts. Three hundred readable words earn full credit, and fewer earn less on a curve that pays the first hundred words most, down to zero for a response with no visible text. If the page only yielded its content after Kuroma fell back to a full browser, whether because the content is rendered by script or because a bot wall stood in the way, it is scored near the bottom, since AI crawlers use no browser.

Our verdict on the evidence

Evidence-backed No major AI crawler executes JavaScript, so content that only appears after scripts run is invisible to AI search. This is among the best-evidenced technical findings in the audit.

Vercel and MERJ AI crawler study (569M GPTBot requests: zero JavaScript execution, no image fetches)

This is the one factor where the score describes what a crawler received rather than what it might prefer: the words counted are the words an AI fetcher gets, and nothing rendered by a client-side component reaches it. The check cannot see which words matter, so a server-rendered page of boilerplate scores as well as a server-rendered article; it can only tell you whether there is anything there at all. A low number here is the one to fix first, because the other factors grade the page as a browser rendered it, and this one says how much of that page an AI crawler was ever handed.

What it is worth

5% of AI Readiness, which is 1.5 points of the overall 100.

How audited pages score on it

Based on 30,738 audited pages.

The median page scores 100, and the middle half of pages score between 57 and 100.

  • Good (75 to 100)21,802 (71%)
  • Moderate (50 to 74)1,635 (5%)
  • Needs work (25 to 49)947 (3%)
  • Critical (0 to 24)6,354 (21%)

24% of audited pages score below 50 on it.

Where it fails most and least

Industries that fail it most

Industries that fail it least

How to fix it

  • Server-render the main content so it is in the HTML response, and treat client-side hydration as an enhancement.
  • Check what curl returns for the page; if the body is a script shell, that is what AI crawlers see.
  • Allow known AI search crawlers through the bot wall so they receive the same HTML a browser does.
  • Keep at least three hundred words of real content in the server-rendered body.

What has changed in this factor

5 versions describe changes that could not be attributed to individual factors, so any of them may have changed this one; they are on the full changelog.

The other factors

All factors, and how the score is built