Free tool · No sign-up
What do AI crawlers actually see on your page?
Most of them do not run JavaScript. If your content is assembled in the browser, they get a blank page — and your site looks perfect to you the whole time.
The problem
Your site looks fine. The crawler sees nothing.
A modern single-page app sends the browser a small HTML shell and a large JavaScript bundle, then builds the page on the visitor's device. Humans see a finished site. A crawler that does not execute JavaScript sees the shell.
We found a live example while auditing a client portfolio: a business operating across eleven US metros, serving a 3.4 KB HTML shell with every word of its content locked inside a 238 KB bundle. To ChatGPT, Claude, Perplexity and AI Overviews that site is effectively blank — for a business whose entire model is local search.
What this can and cannot tell you
An honest limit
This tool fetches your page the way a non-JS crawler does and reports what is present: words, text-to-HTML ratio, headings, structured data, and whether the markup carries the signature of a client-rendered shell.
It does not render the page, so it cannot tell you how much more content JavaScript would add. It tells you what arrives without it — which is the number that matters, because that is what the crawler gets.
AI readability · Questions
What people ask
Do AI crawlers really not run JavaScript?
Most do not. Googlebot renders JavaScript for Search, but the crawlers behind ChatGPT, Claude and Perplexity generally fetch the raw HTML and move on. If your content only exists after hydration, those systems never see it.
My page scored badly. What is the fix?
Server-side rendering or static pre-rendering, so the HTML arrives complete. Most modern frameworks support it — Next.js, Nuxt, Astro and SvelteKit all render on the server when configured to. It is usually a configuration and architecture change rather than a rewrite.
Is a low word count always a problem?
No. A contact page or a short landing page can legitimately have few words. What matters is whether the words you expect are there. Compare what the tool reports against what you see in your own browser — a large gap is the signal, not the raw number.
Does this affect Google rankings too?
Less than it affects AI visibility, because Googlebot does render JavaScript. But rendering is queued and delayed, so client-rendered pages get indexed more slowly and less reliably. Server-rendered HTML is better for both.
What about structured data?
JSON-LD is the most reliable way to state facts an AI can quote — what you do, where you are, what you charge. It sits in the HTML and needs no rendering, so it reaches every crawler. The check reports whether you have any.
Why does the tool not just render the page?
Because then it would be measuring what a browser sees, which you can already check yourself. The point is to show what arrives without JavaScript, since that is what most AI crawlers get. It is also why the tool does not estimate how much content JavaScript adds — it has not run it, so it does not know.
Invisible to AI and not sure why?
Send us the URL. We will tell you whether it is a rendering problem, an indexing problem or an authority problem — they look identical from the outside and the fixes are nothing alike.