Structured data matches the visible text. Schema that contradicts the page reads as unreliable.
The check takes the claims in a page's JSON-LD — the organization name, a product name and description, an article headline and author, each FAQ question and answer — and looks for them in the visible text. Names must match; prose is judged in the full read.
Schema that contradicts the page reads as unreliable. A FAQ answer in the schema that is not on the page, a headline that differs from the H1, a product name no visitor sees — engines learn to distrust the file, and a distrusted file grounds nothing.
For each JSON-LD claim, find it on the page. The quick test: the FAQ questions in the schema and the FAQ questions on the page should be the same list.
Schema & entity check → — free, no signup.
What the agent does: when the agent writes a FAQ block for a page, it writes the FAQPage schema from the same words — the two cannot drift.
Part of the 23 checks · group: Brand identity — JSON-LD on the page, entity links (@id, sameAs), Wikidata record, consistent external profiles.
Names are compared as strings. Whether a long answer in the schema conveys the same fact as the prose needs a model to judge; the free structural check does the string part and marks the rest.
The page check reads your public pages from outside and runs all 22 — free, no signup, no AI engine involved.