Most SEO tools check what Google sees. This checks what GPTBot sees — which is a different thing, on a page that renders client-side or blocks bots at the CDN. Paste any URL, yours or a client's.
Whether GPTBot, ClaudeBot, PerplexityBot and the rest are allowed in at all — robots.txt, bot parity against a normal browser, and whether your content survives without JavaScript. These can cap your whole score, because a page an AI crawler can't fetch scores nothing else.
The machine-readable layer answer engines lean on first: Organization, Product, FAQ and Article schema, and whether what it declares matches what the page shows.
Headings, answerable question formats, readability, and whether a model can lift a clean, quotable passage rather than paraphrasing badly.
Dates, sitemap currency, and the signals that decide whether you're treated as a current source or a stale one.
Some failures make everything else irrelevant. If your robots.txt disallows AI crawlers, no amount of good structured data compensates — the crawler never arrives. So those checks cap the score rather than subtracting from it, and the report says so plainly instead of averaging the problem away.
And when a check can't be run — a firewall blocked us, say — the report tells you that too, rather than scoring it as a pass. How the scoring works.
Anyone can run a scan without an account and see the headline. Signing in unlocks the full report — every check, its evidence, and the fix list — on as many URLs as you like, at no cost.
Selling to product pages instead? Merchant Readiness checks what AI shopping assistants say about your price and stock.