Schema markup, without the guesswork.
Build valid JSON-LD for FAQs, articles, products, organizations, breadcrumbs and how-tos — checked as you type, against the properties each type actually requires. Or paste markup you already have and find out what is wrong with it.
Question and answer pairs — the shape answer engines lift most readily.
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": "FAQPage",
"mainEntity": [
{
"@type": "Question",
"name": "Is there a free plan?",
"acceptedAnswer": {
"@type": "Answer",
"text": "Yes — 20 checks with no card, and nothing converts automatically."
}
}
]
}
</script>Paste this into the <head> of the page it describes. Updates as you type — nothing leaves your browser.
Required properties are present and correctly formatted. That is not the same as a guarantee of a rich result — Google decides that.
- NotevisibilityEvery question and answer here must also be visible to a reader on the page. Marking up content nobody can see is the most common way FAQ markup gets a site penalised — and no tool can check it from the markup alone.
Structured data is you stating a fact instead of hoping it gets inferred.
Everything on your page is prose to a machine until you say otherwise. Schema markup is the one place you get to state — unambiguously, in a format every engine parses the same way — who published this, when, what it is, and what it costs. Nothing else on a page is that unambiguous.
For search engines
It decides rich results — the star ratings, prices and FAQ drop-downs that take up more of a result page than everyone else's blue link.
For answer engines
It removes the inference step. A model that has to work out who wrote a page from a byline in a footer often gets it wrong, or skips the attribution entirely.
For our own audit
Structured data is one of seven groups our page checker scores, worth 20 of 100. Valid markup for the wrong type still fails, which is why this tool checks the type and not just the syntax.
Six mistakes, in roughly the order we find them.
These come from running the page audit against real sites. Every one of them is checked by the validator above — except the last, which nothing can check from markup alone.
Missing @context
Without it the block is plain JSON, not JSON-LD. No engine reads it as schema, and nothing warns you — the page simply has no structured data.
A human-readable date
"21 September 2026" is exactly the thing structured data exists to replace. It has to be ISO 8601: 2026-09-21.
Relative URLs
"/logo.png" means nothing once the markup is read outside your page. Every URL in schema has to be absolute.
An Offer with no currency
price: "49" is not a price. Without priceCurrency it could be dollars, euros or rupees, so the offer is discarded.
Markup that contradicts the page
A Product block claiming a price the page does not show, or an Article dated three years after the copy says. Engines cross-check, and a mismatch costs more than no markup.
FAQ markup for invisible content
Questions that appear in the markup but nowhere a reader can see them. This is the most common way FAQ schema gets a site penalised — and no validator, ours included, can detect it from the markup.
Questions about schema markup
Is this free, and is anything sent to your servers?+
Free, and nothing is sent anywhere. The whole generator runs in your browser — there is no API call, no account, and no record of what you typed. That is also why it updates as you type rather than after a submit.
What replaced Google's Structured Data Testing Tool?+
Google retired it and replaced it with the Rich Results Test, which only reports on the specific rich result types Google supports. This checker is closer to the old tool in spirit: it tells you whether the markup itself is well-formed and complete, including for types Google does not produce a rich result for.
Does schema markup help with AI search, or only Google?+
Both, for different reasons. Google uses it to decide rich results. An answer engine uses it as an unambiguous statement of what a page is — who published it, when, what it sells — which it would otherwise have to infer from prose and often gets wrong. Structured data is one of the seven groups our page audit scores.
Where do I put the markup?+
In the <head> of the page it describes, inside a <script type="application/ld+json"> tag. One page can carry several blocks — an Organization, a BreadcrumbList and an Article together is normal. Put each page's markup on that page, not sitewide.
Will FAQ markup get me a rich result?+
Usually not any more. Google heavily restricted FAQ rich results in 2023 and now shows them mainly for government and health sites. The markup is still worth adding, because answer engines read it as a clean list of question-and-answer pairs — which is the single easiest shape for them to quote.
Can my FAQ answers be hidden behind an accordion?+
Yes, as long as a reader can open and read them. What is not allowed is markup for content that does not appear on the page at all. That is the most common way FAQ markup gets a site penalised, and no validator — ours included — can detect it from the markup alone.
Do I need every property?+
No. Required properties are marked as errors here; recommended ones as warnings. A block with warnings is valid and will be read — the warnings are what stands between valid markup and a rich result.
Can I check markup that is already live?+
Switch to 'Check existing markup' and paste it in. Raw JSON, a script tag, or the whole page's HTML all work — it will find every JSON-LD block, including the @graph shape most CMSes emit.
Markup is one of seven things that decide this.
Structured data is worth 20 of 100 in our page audit. Run the free checker to see the other six groups — crawlability, answerability, content, entity, freshness and speed — on any page you like.
Or read about tracking whether AI names your brand.