Can an AI engine actually read your structured data?
Enter any page. We read its JSON-LD structured data the way a search engine or answer engine does and tell you whether each block can be read at all, what it declares, and whether it carries what an engine needs — plus whether your llms.txt exists and says anything.
A block that looks fine in the browser can be unreadable to every machine: one comment or one stray comma and the parser stops, silently. Agencies ship this fault, and their clients never see it. No signup, no email, no card.
What “cannot be read” means
Structured data is written in JSON, and JSON is unforgiving. A developer who leaves a /* comment */ inside the block to label its sections has produced something a human can read and no machine can. The same goes for a comma after the last item, or single quotes where double quotes are required. The page renders perfectly; the block is dead. Google’s tools report it as an error, but only if someone runs them, and on a site nobody audits that can be never.
What it cannot tell you
- Whether you qualify for rich results. That is Google’s call and Google’s test. We check readability and the fields answer engines need, not eligibility.
- Whether the facts are true. A block can be perfectly readable and say your business is open on days it is closed. We read what it declares; we do not check it against the world.
- Whether the engines name you. Readable structured data is the floor, not the result. Whether ChatGPT or Perplexity recommend you when a customer asks is a separate measurement, and it is the one the report makes.
Frequently asked questions
What is structured data, and why does an AI engine care?
Structured data is a block of machine-readable facts hidden in a page’s code, in a format called JSON-LD, that states plainly what the page is about: the business name, address, services, the questions it answers. Search engines and AI answer engines read it to describe you without guessing. When the block does not parse, they read nothing from it at all.
My site has structured data. How can it be unreadable?
JSON is strict. A comment left inside the block, a trailing comma, single quotes instead of double quotes — any one of these makes a parser stop, and it stops silently. The page still looks fine in a browser, which is why the fault survives. This tool reports the block as unreadable and says which of those faults it found.
Does this tool tell me if I qualify for rich results on Google?
No. Google’s Rich Results Test is the authority on that and we do not pretend otherwise. This tool answers a narrower question: can the block be read at all, what types does it declare, and does it carry the fields an answer engine needs — a name, a locality, linked profiles, FAQ answers that also appear on the page.
What is llms.txt, and should I have one?
A plain text file at yoursite.com/llms.txt that describes the site for language models. It is optional. Google has said it does not use the file, and a large study found most llms.txt files are never requested by anything. We check it because a file that exists and is empty is worse than none — it is a claim of readiness with nothing behind it.
Why do you check whether FAQ answers are visible on the page?
Because schema that says more than the page does is the pattern search engines penalise, and it is the easiest one to slip into. If a question exists in the code but a visitor cannot find it on the page, it is markup describing content that is not there.
Is this tool really free?
Yes. No signup, no email, no card, and the whole result is on screen. It reads one page, once, the way a browser would. We sell paid reports that answer different questions: whether AI engines name your business for real buyer searches, and what SEO work is actually showing up on your site.
Last reviewed 2026-09-14.