Free tools · 01
Schema Markup Checker
Paste any URL. See exactly which structured data types are on the page, which are missing, and get the JSON-LD to paste.
Check a page
224 pages checked so far
One page, six types, no account. We fetch the raw HTML exactly as GPTBot receives it and read only the markup that validates.
Example result · everlane.com
ExampleMissing 4 of 6: Product, FAQPage, BreadcrumbList, Article.
What is schema markup?
Schema markup is a small block of code that states what a page is, in a vocabulary machines already agree on. A product page can say it is a Product with a price and a brand. A post can say it is an Article with an author and a date. A support page can say it is an FAQPage with questions and answers attached.
It matters because a model reading your page has to infer all of that from prose unless you label it. Labelled pages are cheaper to quote and harder to misread, which is why schema is the dimension where the cohort splits hardest.
How we check it
- 01We fetch the raw HTML with no JavaScript executed, because that is what GPTBot does. If your markup is injected on the client, GPTBot never sees it and neither do we.
- 02We parse all three formats: JSON-LD script blocks, microdata attributes and RDFa. They are three ways of writing the same vocabulary and we treat them equally.
- 03We only count markup that actually validates. A block missing a required property, or malformed JSON, is reported as missing rather than credited.
What is a good schema score?
Across the 20 brands in the current CLIC Index the schema dimension averages 34 out of 100. 1 of the 20 scored a flat zero: not low, zero, with no valid structured data found anywhere we could reach. Anything above 34 puts you ahead of the cohort average, and above 70 puts you at the top of it.
34
Cohort schema average
1
Of 20 scored a flat zero
40
Point gap, top seven vs bottom seven
Figures from The CLIC Index, recomputed from the published rows.
How to improve it
- 01Add Organization and WebSite site-wide. One JSON-LD block in the layout, name, URL, logo, social profiles. This is the cheapest point on the board.
- 02Label each template for what it is. Product on product pages with price and availability, Article on posts with author and date, BreadcrumbList where there is a real hierarchy.
- 03Mark up questions you already answer. If a support page has questions and answers in prose, FAQPage costs you nothing new to add.
- 04Render it server side. If the block only exists after hydration, no AI crawler will read it.
- 05Re-check the page after deploying. Validation failures are the most common reason a block that exists still does not count.
Frequently asked questions
Is this checker really free?
Yes, and there is no account. It runs on the crawler we already operate for the full scan, so a single page costs us close to nothing. We do not rate limit it for normal use and we do not email you unless you ask us to.
Does schema help with ChatGPT and Claude?
It helps them understand the page, which is not the same as guaranteeing a citation. Structured data states plainly what a page is, what you sell, who wrote it and what questions it answers, in a format that survives having all the design stripped away. In the current index schema is the single dimension that separates the top of the cohort from the bottom by the widest margin.
Why does my schema not show up in your result?
Almost always because it is injected by JavaScript after the page loads. We fetch the raw HTML with no JavaScript executed, because that is what GPTBot does. The other common cause is markup that does not validate: a missing required property or a malformed JSON-LD block. We only count markup that actually validates.
What is the difference between JSON-LD, microdata and RDFa?
They are three ways of writing the same vocabulary. JSON-LD is a script block in the head or body and is the format we recommend, because it is separate from your markup and easy to change. Microdata and RDFa are attributes added to the HTML itself. We parse all three and treat them equally when scoring.
How many schema types should one page have?
Usually two to four. A site-wide Organization and WebSite, then whatever the page actually is: Product on a product page, Article on a post, FAQPage where you genuinely answer questions, BreadcrumbList where there is a real hierarchy. More than that and you are usually describing things the page does not contain, which is worse than having none.
Other free tools
Head to head compare
Your site against three you lose to, same crawler, same minute.
Returns: 5 DIMENSIONS × 4 DOMAINS
Run a comparison →FAQ schema builder
Turn a list of customer questions into valid FAQPage markup.
Returns: JSON-LD BLOCK
Join the waitlist →Robots and crawler check
Whether GPTBot, ClaudeBot and PerplexityBot are allowed to read your site at all.
Returns: 6 CRAWLERS
Join the waitlist →