Skip to main content Scroll Top

Schema Markup Validator: Does Your Page Support Its Structured Data?

Updated September 2026 · Written and maintained by the Progression Agency strategy team

A schema markup validator that checks the half the others skip. Syntax checkers confirm your structured data is valid; this confirms your page actually says what it claims.

On this page · 11 sections
  1. Valid is not the same as true
  2. What it skips, and why
  3. The three mismatches we find most often
  4. Fix the copy or fix the markup — not neither
  5. How to get clean inputs
  6. Why plugins produce most of these mismatches
  7. What this tool will not tell you
  8. How this differs from every other schema markup validator
  9. The reference tables, in one place
  10. Everything else we have written on search, AI and getting found
  11. A replacement for the retired Structured Data Testing Tool

The short answerStructured data is a description of what is on a page. Every existing checker confirms the description is well-formed; none of them look at the page to see whether it is accurate. Paste your JSON-LD and the visible copy of the same page and this tool lists every value the copy does not support, skipping the twenty or so machine-facing properties that are never expected in prose. The three mismatches we find most often are FAQ answers that exist only in markup, ratings with no visible reviews, and prices that appear nowhere on the page.

This does not check syntax, required properties or rich-result eligibility. Run a syntax validator as well; the two answer different halves of the question.

Schema versus visible copy validator

Paste your JSON-LD and the visible text of the same page. Every value in the markup that does not appear in the copy is listed, because markup describing things the page does not say is the most common reason structured data is ignored.

Results appear here.

Runs entirely in your browser. Nothing you paste is uploaded, logged or stored — which is also why it takes pasted text rather than a URL: a browser cannot read another site’s files without that site’s permission.

Valid is not the same as true

Every existing checker confirms your markup is well-formed; none of them look at your page to see whether it is accurate.

Every existing checker answers the easy question

Syntax validators confirm that your JSON-LD parses, that the type you declared has the properties it needs, and that the result is eligible for a particular presentation. All of that is useful and none of it involves looking at your page. A validator cannot know whether the rating you declared corresponds to reviews a visitor can read, because it never sees the visitor’s view.

The hard question is whether the page supports the markup

Structured data is a description of what is on a page. When the markup asserts something the page does not — a price that appears nowhere, an answer that exists only in the markup, a review count with no reviews — the description has stopped describing. At the mild end that is why markup gets ignored. At the serious end, published guidance treats marked-up content that is not visible to users as a violation, and the consequences are manual rather than algorithmic.

So this tool only asks that one question

It walks every value in your markup, skips the properties that are never expected to appear in visible text, and reports what is left that the copy does not support. It does not check your syntax and it does not tell you whether you are eligible for anything. Run a syntax validator as well; they answer different halves.

Why the gap matters more than the syntaxWhy the gap matters more than the syntax
Two different questions. Is the markup valid, and is the markup honest. Everything already checks the first; almost nothing checks the second.

What it skips, and why

Roughly twenty machine-facing properties are set aside by design, because reporting them would bury the real findings in noise.

Machine-facing properties

URLs, image paths, identifiers, language codes, context declarations and the structural plumbing of a graph are never expected to appear in readable copy. Reporting them would bury the real findings under noise, so roughly twenty property names are skipped by design.

Dates in machine format

A page saying ‘updated September 2026’ supports a dateModified of 2026-09-17 perfectly well, and demanding an ISO string in the prose would be absurd. Dates are counted and skipped, and the count is shown so you know how many were set aside.

Very short values

Two-character values and bare numbers match almost anything, so including them would produce false passes rather than useful findings. They are ignored.

Paraphrase is tolerated

Longer values are matched on words rather than exact strings, so a description that is worded slightly differently on the page still counts as supported. That deliberately errs toward not crying wolf; a value reported as missing is usually genuinely missing.

Valid — Not the same as true. Syntax passes regardless.
Ratings — The costly case. Guidance is explicit.
Prices — Commonly mismatched. Misleads before the click.
FAQ — Most common of all. Answers only in markup.
Dates — Expected to differ. Skipped by the tool.
URLs — Expected to differ. Skipped by the tool.

The three mismatches we find most often

Three patterns account for the great majority of what this check surfaces, and none of them are typed deliberately.

FAQ answers that exist only in the markup

By a wide margin the most common. A page has an FAQ block in its structured data with eight questions and answers, and the visible page has four of them, or none. This usually happens because an SEO plugin offers an FAQ field that writes markup without writing content, and nobody noticed the two had drifted apart.

Ratings and review counts with nothing visible

The costly one. Aggregate rating markup on a page with no reviews on it is explicitly addressed in published guidance, and it is the category most associated with manual action rather than quiet demotion. It is also easy to introduce by accident — a theme setting, a plugin default, a template copied from another site.

Prices that appear nowhere on the page

Common on service sites that publish Offer markup because a plugin asked for a number, while the page itself says ‘contact us for pricing’. The markup then advertises a figure the reader never sees, which is misleading in exactly the direction that gets noticed.

Where this fits with the rest

Markup is one of several things that has to agree. Our audit checklist covers the others, including the entity records that sit outside your site entirely.

Which mismatches matter, and how muchWhich mismatches matter, and how much
Green rows are expected mismatches and the tool skips them. Red rows are the ones that make markup untrustworthy, and they are what it reports.

Fix the copy or fix the markup — not neither

Every row in the output is a decision between two valid fixes, and leaving the gap is not one of them.

If the value is true, put it on the page

A price you actually charge, an answer you actually give, a credential you actually hold — these belong in the visible copy, where they help a reader and can be extracted. Moving them from markup into prose is almost always the better fix, because prose works for every consumer and markup works only for the ones that read it.

If the value is not true, take it out

Inherited template values, placeholder ratings and prices from a page the template was copied from should be deleted rather than justified. Nothing is gained by markup describing a page that does not exist.

What we would not do

We would not add invisible text to make the markup match. Hidden content added to satisfy structured data is a worse version of the original problem and is addressed directly in published guidance.

Paste — The JSON-LD. Script contents only.
Paste — The visible copy. What a reader sees.
Walk — Every scalar value. Nested objects included.
Skip — Machine properties. Twenty of them.
Match — On words. Not exact strings.
Report — What is unsupported. Property and value.

How to get clean inputs

The JSON from the script block and the rendered text a reader sees — pasting HTML into the copy box defeats the check entirely.

Getting the JSON-LD

View source, or open the page’s structured data through any inspector, and copy the contents of the script block — the JSON itself, without the surrounding tags. If a page has several blocks you can paste them one after another and the tool will read each.

Getting the visible copy

Select the page in a browser, copy, and paste. That gives you what a reader sees, which is exactly the comparison basis you want. Do not paste the HTML: markup inside HTML would count as visible text and defeat the check.

A note on JavaScript-rendered pages

If content appears only after scripts run, copy the page as rendered rather than from view source, or the comparison will report copy that exists as missing. That itself is worth knowing about, and it is what the access side of things and a rendering check are for.

How costly each mismatch class isHow costly each mismatch class is
The top row is the one that attracts manual action. The bottom row is what the tool deliberately tolerates by matching on words rather than on exact strings.

Why plugins produce most of these mismatches

Two mechanisms account for almost all of them, and neither involves anybody deciding to mislead.

A required field with no content behind it

The mechanism is almost always the same. A plugin exposes a field — rating, price, FAQ answer — because the schema type it is emitting has that property. Somebody fills it in once, during setup, because the field is there. Nothing ever writes the corresponding content onto the page, and nothing ever checks.

Templates carry values between pages

The second mechanism is duplication. A template built for one page is reused for thirty, and the values baked into it travel with it. Multi-location sites produce this constantly: one address, one phone number and one set of opening hours marked up across every location page.

Which is why this is a maintenance check, not a one-off

Neither mechanism involves anybody deciding to mislead, and neither leaves a trace in a changelog. Running the comparison after template changes, plugin updates and bulk edits catches them at the point they are introduced rather than months later.

Validators — Syntax and eligibility. Run one too.
Guidance — Published policies. The authority.
Your page — The visible copy. The thing described.
Your markup — The description. Must match.
Rich results — One outcome. Not the only one.
Entity records — Off-site. Also must agree.

Want this done for your site?We build and maintain the search, content and paid programmes described on this page.

Get a free proposal

What this tool will not tell you

Three things a clean comparison does not establish, stated so a clean result here is not mistaken for a clean result everywhere.

Whether your markup is syntactically correct

It does not parse-check beyond what it needs to read your values, and it will happily compare markup that a syntax validator would reject. The two checks are complements and we would rather be explicit about the boundary than let a clean result here imply a clean result everywhere.

Whether you will get a rich result

Eligibility depends on type requirements, quality signals and presentation decisions that no browser-side tool can evaluate. A page can pass this comparison completely and still never show an enhanced result.

Whether the values are actually true

It checks that your page says the same thing as your markup. If both say something untrue, both agree and the tool reports nothing. Consistency is the property being measured, and consistency is not the same as accuracy.

What each kind of checker can seeWhat each kind of checker can see
They are complements, not alternatives. Run a syntax validator for the first three columns and this for the fourth.

How this differs from every other schema markup validator

Every structured data checker on the market answers the same question. This one answers the question none of them ask.

What a conventional schema markup validator does

It parses your JSON-LD, confirms the type you declared carries the properties it requires, and tells you whether the result is eligible for a particular presentation. The official testing tools do this, they are free, and they are worth running.

What no structured data checker checks

Whether your page says what your markup says. A structured data checker never loads the rendered page, so aggregate rating markup on a page with no visible reviews, a price that appears nowhere, or an FAQ answer that exists only in the markup all pass cleanly.

Why that is the half that matters

Published guidance treats marked-up content that is not visible to users as a violation, and the consequences there are manual rather than algorithmic. A clean syntax report on markup the page does not support is a false reassurance, which is why this exists as a separate check.

Run both, in this order

A conventional schema markup validator first, so you know the markup is well-formed. This second, so you know it is honest. Together they take about three minutes.

Rating with no reviews — Policy problem. Not a grey area.
Price only in markup — Misleading. Fix one or the other.
Invisible FAQ answers — Very common. Put them on the page.
Copy-pasted markup — Risky. Values belong to another page.
Markup as a shortcut — Wrong model. It describes, it does not assert.
Ignoring the gap — The default. And the reason markup is ignored.

The reference tables, in one place

Four tables follow: what the tool skips, the mismatch classes ranked by severity, what to do with each finding, and how this differs from a syntax validator. They summarise the sections above and are meant to be usable on their own.

Properties the tool deliberately skips
Property classExamplesWhy it is skipped
Context and typing@context, @type, @idStructural, never in prose
Linksurl, sameAs, mainEntityOfPageNever expected in visible text
Mediaimage, logo, contentUrl, thumbnailUrlPaths, not prose
Graph plumbing@graph, isPartOf, publisher, breadcrumbStructure, not content
Machine datesdatePublished, dateModified, validThroughISO format, counted separately
Measurementswidth, height, positionNumeric metadata
Currency codespriceCurrencyA code, not copy
Mismatch classes, ranked
MismatchSeverityWhyUsual cause
Rating or review count with nothing visibleHighestAddressed directly in published guidanceTheme or plugin default
Price that appears nowhere on the pageHighMisleads before the clickOffer markup with a required field
FAQ answers absent from the copyHighThe most common mismatch we seePlugin field that writes markup only
Availability or stock claimsHighCheckable and frequently staleFeed or template default
Name or address differing from the pageModerateFragments the entityMulti-location templates
Credentials or awards not mentionedModerateUnsupported assertionCopied template
Loosely paraphrased descriptionLowUsually fine, and tolerated by the matcherNormal editing
What to do with each finding
FindingIf the value is trueIf the value is not true
Rating or reviewsPublish the reviews on the pageRemove the markup
PriceState it in the copy, with a dateRemove the Offer block
FAQ answerWrite the answer into the pageRemove the question
AvailabilitySay it on the pageRemove or correct it
AddressCorrect the page or the markup so both agreeRemove the wrong one
Award or credentialMention it in the copyRemove it
DescriptionUsually no actionRewrite to match the page
This tool against a syntax validator
QuestionSyntax validatorThis tool
Does the JSON parseYesNo
Are required properties presentYesNo
Is the type used correctlyYesNo
Am I eligible for a rich resultYesNo
Does the page support these valuesNoYes
Is anything asserted that is not visibleNoYes
Should I run bothYesYes
Structured data problems by how they are caughtStructured data problems by how they are caught
Everything in the left column is already caught by free tooling. The right column is where the expensive problems live and where almost nothing looks.
How to work through the outputHow to work through the output
Step four is the actual decision. A value that should be true belongs in the copy; a value that is not true belongs out of the markup.
What the comparison coversWhat the comparison covers
The skipped properties are URLs, images, identifiers and machine formats, which are never expected to appear in visible copy.
Compare — Markup to copy. Both pasted.
Separate — Expected from real. Flags do this.
Decide — Copy or markup. Per row.
Fix — One of the two. Never neither.
Re-run — Until clean. Or deliberate.
Validate — Syntax separately. Different question.
Supported — Every value. By visible text.
Honest — Above valid. Syntax is the easy half.
Specific — Property-level. Not a score.
Tolerant — Of paraphrase. Word matching.
Local — In your browser. Nothing uploaded.
Complementary — To validators. Run both.

Everything else we have written on search, AI and getting found

A replacement for the retired Structured Data Testing Tool

Google retired the Structured Data Testing Tool and replaced it with the Rich Results Test, which answers a narrower question: whether a specific rich result is possible.

That left a gap. The old tool would parse and show you everything on a page. The Rich Results Test only reports on types it supports, so markup that is present but unsupported simply does not appear, and people read that silence as an error.

What replaced it, and what each replacement answers

  • Rich Results Test: whether Google could show a supported rich result. Narrow by design.
  • Schema Markup Validator at validator.schema.org: whether the markup is valid schema.org at all.
  • This tool: whether the markup agrees with the words a visitor actually sees, which neither of the others checks.

Why the third question is the one that gets sites penalised

Both Google tools check syntax and eligibility. Neither compares your markup against your visible content. Structured data describing content that is not on the page is a policy violation, and the penalty applies at site level rather than to the offending block, which is why it is worth checking deliberately rather than assuming.

What to run, and in what order

  1. Validate the syntax at validator.schema.org so you know the markup parses.
  2. Run the Rich Results Test to see whether a supported enhancement is possible.
  3. Run this checker to confirm the markup does not claim anything the page does not say.
  4. Fix contradictions before chasing eligibility, because a contradiction is the more serious problem.
The retired tool, and what now answers each of its jobs
Job the old tool didWhat answers it nowWhere it falls short
Parse all structured data on a pageSchema Markup ValidatorDoes not report Google eligibility
Report syntax errorsSchema Markup ValidatorDoes not judge content accuracy
Show unsupported typesSchema Markup ValidatorGoogle’s own tool hides them entirely
Check rich result eligibilityRich Results TestOnly for supported types
Check markup matches the pageNothing from GoogleThis is the gap this tool fills

AI, AEO and what is changing

Websites and design

Choosing and working with an agency

Social, content and brand

By industry and by situation

Want the markup and the copy reconciled properly

We audit structured data against the page, the entity records off your site and the content itself, and hand back the specific edits rather than a score.

/contact

Not sure which of these applies to you?Tell us the situation and we will say plainly what we would do first, and what we would not.

Talk it through

Getting found in search

AI, AEO and what is changing

Paid media and lead generation

Websites and design

Choosing and working with an agency

Social, content and brand

By industry and by situation

Frequently asked questions

Is this a schema markup validator?
It validates the half no other schema markup validator covers: whether the visible copy of your page supports the values in your structured data. It does not parse-check syntax, so run a conventional validator as well.
How is this different from a structured data checker?
A structured data checker confirms your JSON-LD is well-formed and carries the required properties. It never looks at your page, so markup asserting things the page does not say passes it cleanly. This tool only checks that relationship.
Which schema markup validator should I run alongside it?
Any conventional one; the official testing tools are free and adequate. Run that first to confirm the markup parses, then run this to confirm the page supports it.
What replaced the Structured Data Testing Tool?
Google retired it in favour of the Rich Results Test, which reports only on whether a supported rich result is possible. The Schema Markup Validator at validator.schema.org took over general schema.org validation.
Why does the Rich Results Test not show all my markup?
Because it only reports on types it supports. Valid markup for an unsupported type simply does not appear, and that silence is frequently misread as an error.
Is there still a structured data testing tool?
Yes, in two halves. validator.schema.org checks whether markup is valid schema.org; the Rich Results Test checks Google eligibility. Neither checks whether your markup agrees with your visible content, which is what this tool does.
What is the difference between the Rich Results Test and the Schema Markup Validator?
The Rich Results Test answers whether Google could display a supported enhancement. The Schema Markup Validator answers whether the markup is valid schema.org at all. They answer different questions and both are worth running.
Why check markup against the page copy?
Because structured data must describe content a visitor can see. Marking up content that is not on the page is a policy violation and the consequence applies at site level. Neither Google tool checks this.
What does this tool actually check?
Whether every value in your structured data also appears in the visible copy of the same page. It does not check syntax, required properties or rich-result eligibility — a syntax validator already does those and you should run one as well.
Why does the gap matter?
Because structured data is a description of what is on a page. When markup asserts things the page does not say, it has stopped describing. Mildly, that is why markup gets ignored; seriously, published guidance treats marked-up content that is not visible to users as a violation.
What is the most common mismatch?
FAQ answers that exist only in the markup. It usually happens because an SEO plugin offers an FAQ field that writes structured data without writing page content, and the two drift apart.
What is the most serious mismatch?
Aggregate rating or review-count markup on a page with no visible reviews. It is addressed explicitly in published guidance and is the category most associated with manual action rather than quiet demotion.
Why are dates skipped?
Because a page saying ‘updated September 2026’ supports an ISO dateModified perfectly well, and requiring the machine format in prose would be absurd. Dates are counted so you can see how many were set aside.
Why are URLs and images skipped?
They are machine-facing values that are never expected to appear in readable copy. Reporting them would bury the real findings in noise.
Does it require an exact text match?
No. Longer values are matched on words rather than exact strings, so a description worded slightly differently on the page still counts as supported. The matcher deliberately errs toward not crying wolf.
What should I paste into the copy box?
Select the page in a browser, copy and paste. That is what a reader sees, which is the right comparison basis. Do not paste HTML — markup inside it would count as visible text and defeat the check.
What if my content only appears after JavaScript runs?
Copy the page as rendered rather than from view source, or the comparison will report copy that exists as missing. That situation is itself worth investigating separately.
Can I paste several JSON-LD blocks?
Yes. Paste them one after another and the tool reads each, reporting across all of them together.
Should I add hidden text so the markup matches?
No. Hidden content added to satisfy structured data is a worse version of the original problem and is addressed directly in published guidance. Either put the value in visible copy or remove it from the markup.
Which fix is usually better, copy or markup?
Putting the value in the copy, when it is true. Prose works for every consumer; markup works only for the ones that read it.
Where do these bad values usually come from?
Theme defaults, plugin fields with required inputs, and templates copied from another site. Very few are typed deliberately, which is why nobody notices them.
Does it upload my page?
No. Both inputs are pasted and everything runs in your browser; the page you are checking is never uploaded. The only thing sent to us is the contact detail you enter to unlock the full findings.
Does clean output mean my structured data is correct?
It means every value is supported by your copy. It says nothing about syntax, required properties or eligibility, which is why the page tells you to run a syntax validator as well.
What if a value is reported missing but is genuinely on the page?
Check that you pasted the rendered copy rather than a partial selection. If it still reports, the wording may differ enough that word matching fails, in which case you can treat the row as deliberate.
Does this apply to AI answers as well as search?
The same principle applies more strongly. Values that exist only in markup are not what a reader or a retrieval system sees quoted, so the visible copy is doing the work either way.
How often should I run it?
After any template change, plugin update or bulk content edit — those are the three events that introduce mismatches without anybody deciding to.
Can I use it on client sites?
Yes, with no restriction and no attribution requirement.
Does it handle nested objects and graphs?
Yes. It walks the full structure including nested objects and arrays, skipping structural keys rather than stopping at the top level.
Why not fetch the page automatically?
A browser cannot read another origin’s pages without permission, and proxying your site through our server to save a copy and paste is a worse trade than it sounds.
What should I do first if the table is long?
Sort by severity in your head: ratings, then prices, then FAQ answers. Those three account for most of what matters and most of what gets noticed.

Want this done for your site?We build and maintain the search, content and paid programmes described on this page.

Get a free proposal

Get a free marketing proposal

Tell us what you are trying to grow and we will come back with a plan, not a pitch deck. Same-day reply on weekdays.

Our website uses cookies from third party services to improve your browsing experience. Read more about this and how you can control cookies by clicking "Privacy Preferences".
Privacy Preferences
When you visit our website, it may store information through your browser from specific services, usually in form of cookies. Here you can change your privacy preferences. Please note that blocking some types of cookies may impact your experience on our website and the services we offer.
Privacy Policy
You have read and agreed to our privacy policy
Required
Contact Us