Structured data checker

Every script of type application/ld+json on the page is parsed. The ones that do not parse are listed with the error, because a block that does not parse is a block Google never reads. Every typed node in the ones that do, @graph and nested nodes included, is checked against the fields Google documents for its rich result: Article (headline, image, dates, author), Product (name, and offers, a review or a rating), SoftwareApplication (name, category, operating system, offers or rating), FAQPage (questions with accepted answers), HowTo (name, steps), BreadcrumbList (items in order with names), Organization (name, url, logo), WebSite, Review, Event, LocalBusiness, VideoObject, Person and a few more. For each node the table says which rich result it can earn and whether it has what that needs, what is required and missing, what is recommended and missing, and what changed: FAQ rich results have been limited to government and health sites since August 2023, and How-to rich results were retired in September 2023. Microdata nodes are counted with their types. And for each node, a mended copy: its own values kept, the missing fields filled with placeholders to replace, ready to paste as a script tag.

Structured data is the one place a page says what it is in a form every reader agrees on: this is an article, published then, by them; these are questions and their answers; this is a product at this price. Google uses some of it for rich results, the extra lines under a result, and all of it to understand the page. Assistants read it too. A block with a syntax error or a missing required field is worth exactly nothing, and no browser will tell you.

Well-formed is not the same as shown. Google decides whether a page gets a rich result, on its own rules, and says which fields it wants without promising to use them. The checker tells you the block is right; it cannot tell you what Google will do with it.

• The verdict says whether every block parses, how many types are declared, and whether any node lacks a required field. A page with no structured data is not an error.
• Rich result names what the type can earn and whether the node is in shape for it: in shape, not yet (a required field is missing), or no longer shown (FAQ and How-to).
• Required and missing are the fields Google lists as required; without them the rich result is not shown. Recommended and missing are the ones Google says improve it.
• Note carries the specifics: a FAQPage entry that is not a Question with an answer, a breadcrumb out of order, an image as one string where Google prefers an array, a software app without an operating system.
• Mended JSON-LD opens a copy of the node with the gaps filled by placeholders. Replace the placeholders with what is on the page; never with what is not.

JSON-LD is judged; microdata is counted, not judged; RDFa is not read. Pages that inject their markup with a script after load show none here. The rules cover the types a product site carries; a type not listed is shown, not judged. And shape is not eligibility: Google decides whether to show a rich result on its own rules, and the Rich Results Test in Search Console is the word on what it would show for a page.