Schema Validator

Check the schema markup on any page. See the errors, and what your structured data tells AI about your brand.

or to check it before it goes live

It's free to run

What is schema markup?

Schema markup is code that states the facts of a page in a form a machine can read without guessing. A person sees a name, a price and five stars. The markup says: this is a product, this is its name, this is the price, this is the rating.

The vocabulary comes from schema.org. Structured data is the general name for the same thing, and JSON-LD is the format most sites use to write it: one block of code in the page, separate from what visitors see.

Search engines use it to understand a page and to show richer results. AI systems read the same block when they fetch the page. That makes it one of the few places where you can say plainly who you are and what you offer, in words a machine cannot misread.

What this schema validator checks

That the code parses. One stray comma makes a whole JSON-LD block unreadable, and a crawler then ignores everything in it. The validator names the block and the line where it breaks.

That every item says what it is. An item needs a @type, and the block needs a @context that points at schema.org. Types are case-sensitive, and the validator catches the ones written in the wrong case.

That each type has what it needs. A Product without an offer, a FAQ without answers, a video without an upload date. Missing required properties are errors. Missing recommended ones are listed apart, so you can tell what is broken from what could be better.

That the values make sense. Empty properties, links that are not full URLs, dates that are not dates, prices with a currency sign inside the number.

What the markup says about your brand. The validator pulls out the Organization item and shows it as a machine reads it: name, description, site, logo and the profiles listed in sameAs. This is the schema markup checker part most tools leave out.

Code you paste is checked in your browser and never leaves the page.

How to validate schema markup

  1. Enter the page address. Use the page itself, not only the homepage. Each page carries its own markup.
  2. Read the lines at the top. They say whether schema was found, whether it has errors, and whether it tells a machine who runs the site.
  3. Look at the brand table. It shows your Organization data the way a machine gets it. Anything marked missing is a gap.
  4. Go through the items. Each schema item on the page is listed with its type and its result.
  5. Fix what is under Validation. Errors come first, then warnings, then recommended properties.
  6. Test the fix before you publish. Paste the new JSON-LD into the validator and check that the errors are gone.

What schema markup tells AI about your brand

When an assistant answers a question about a company, it works from what it could read. Prose leaves gaps. Is this the company or a reseller? Is that the official site? Do these two profiles belong to the same brand? Schema markup answers those questions in one place.

Organization is the item that matters most

An Organization item on the homepage states your name, your URL and your logo. Add a description and a machine has your own one-line summary of what you do, in your words.

sameAs connects the dots

sameAs lists your other profiles: social networks, directories, review sites. It tells a machine that all of them are one entity. A brand whose profiles are tied together is easier to recognise than one scattered across the web under slightly different names.

One name everywhere

If the Organization is called one thing and the WebSite another, a machine has to decide which is right. The validator flags the mismatch. Pick one name and list the other as alternateName.

Is structured data used by ChatGPT?

Here is what is known. Google and Microsoft both say structured data helps their systems understand a page. OpenAI, Anthropic and Perplexity have not said how much weight their models give it. Their crawlers do fetch the HTML, and the JSON-LD is part of it.

So schema markup is not a switch that makes assistants cite you. It removes doubt about who you are. For one block of code, that is worth it.

Markup added by a script may never be seen

Most AI crawlers do not run JavaScript. If a tag manager adds your schema after the page loads, the markup is not in what they fetch. This validator reads the same HTML, so if it finds nothing on a page you know has schema, that is the reason. The Agent Readiness Checker shows how much of a page is there before any script runs.

Schema types this validator checks

These are the types with rules behind them. Other schema.org types are read and listed, with no verdict on their properties.

TypeWhat it describesRequiredRecommended
OrganizationWho is behind the sitename, urllogo, sameAs, description
LocalBusinessA business with a physical locationname, addressurl, telephone, image, opening hours, sameAs
WebSiteWhat the site is calledname, urlNone
PersonAn author or a foundernameurl, sameAs, jobTitle
ArticleA post or a news storyheadlineauthor, datePublished, dateModified, image, publisher
ProductSomething you sellname, and one of offers, review or aggregateRatingimage, description, brand, sku
OfferThe price of a productpriceCurrency, and price or priceSpecificationavailability
SoftwareApplicationAn app or an online toolname, and one of offers, aggregateRating or reviewapplicationCategory, operatingSystem, description
FAQPageQuestions with their answersmainEntity, with a name and an acceptedAnswer for each questionNone
BreadcrumbListWhere the page sits in the siteitemListElement, with a position for each itemNone
VideoObjectA video on the pagename, description, thumbnailUrl, uploadDate, and contentUrl or embedUrlduration
EventSomething happening on a datename, startDate, locationendDate, image, description, offers, organizer
RecipeA recipename, imagerecipeIngredient, recipeInstructions, author, totalTime
HowToStep by step instructionsname, stepimage, totalTime
JobPostingAn open positiontitle, description, datePosted, hiringOrganization, jobLocationvalidThrough, employmentType, baseSalary
Review and AggregateRatingRatingsratingValue, plus an author for a review or a count for an aggregatedatePublished

Subtypes follow their parent. A Restaurant or a Dentist is checked as a LocalBusiness, a BlogPosting or a NewsArticle as an Article, a WebApplication as a SoftwareApplication.

Common schema markup errors

A comma after the last property. JSON does not allow it. The whole block stops being readable.

The wrong case in a type. organization is not Organization. Types and property names are case-sensitive.

No @context. Without it the type names are only words.

A Product with no offer. A product needs a price, a review or a rating to be useful to anyone reading it.

A price with a currency sign. "price": "$19" is text. Write "price": "19.00" and put USD in priceCurrency.

Relative links. /logo.png means nothing outside your site. Use the full address.

Dates in a local format. Write the year first, as in 2025-03-14, not 14/03/2025.

Markup for things the page does not show. Ratings and reviews in the code that no visitor can see break Google's rules.

Two plugins writing the same item. The page ends up with two Organization items that disagree. Keep one.

Schema validator vs Rich Results Test

The tools answer different questions, and it is worth using more than one.

ToolThe question it answersWhat it leaves out
Google's Rich Results TestCan this page get a rich result in Google Search?Types Google does not use for rich results
The Schema.org validatorDoes the markup follow the schema.org vocabulary?Whether anything required for search is missing
This schema validatorIs the markup correct, and what does it tell a machine about the brand?Google's eligibility rules for each rich result

If your goal is a rich result in Google, finish with Google's own test. If your goal is to be understood by search engines and AI systems alike, start here.

JSON-LD, microdata and RDFa

Schema markup can be written in three formats. They carry the same vocabulary.

JSON-LD is a block of code in a script tag. It is separate from the visible page, which makes it the easiest to write and to maintain. Google recommends it, and most sites use it.

Microdata puts attributes such as itemscope and itemprop on the HTML elements themselves. It works, and it breaks easily when a template changes.

RDFa does the same with a different set of attributes and is the least common of the three.

This validator checks JSON-LD in depth. For microdata it lists the types it finds on the page, so you know they are there.

Frequently asked questions

What is a schema validator?

A schema validator reads the structured data on a page and tells you whether it is written correctly. It checks that the code parses, that each item names its type, and that the properties each type needs are there. Schema checker and structured data testing tool are other names for the same thing.

How do I check the schema markup on my website?

Enter the page address in the validator above. It reads the page the way a crawler does, lists every schema item it finds and marks what is missing or wrong. To test code before you publish it, paste the JSON-LD instead.

Why does the validator find no schema on my page?

Usually because a script adds the markup after the page loads. A tag manager or a client-rendered app does this. The validator reads the HTML the server sends, which is also what most AI crawlers get. Put the JSON-LD in the page source and it will be found.

Does schema markup help with ChatGPT and other AI assistants?

It helps a machine read your page without guessing. Google and Microsoft both say structured data helps their systems understand content. OpenAI, Anthropic and Perplexity have not said how much weight their models give it. Treat it as a way to state the facts about your brand clearly, not as a switch that makes assistants cite you.

Which schema types matter most for a brand?

Organization comes first, on the homepage, with your name, URL, logo and sameAs links. WebSite gives the site its name. After that it depends on the page: Product for what you sell, Article for what you publish, FAQPage for questions and answers, LocalBusiness if customers visit you.

What is the difference between this and Google's Rich Results Test?

Google's test answers one question: can this page get a rich result in Google Search. It only looks at the types Google supports for that. This validator checks the markup itself and shows what it tells any machine about your brand, including types Google does not use for rich results.

Is JSON-LD better than microdata?

For most sites, yes. JSON-LD sits in one block apart from the visible HTML, so it is easier to write, to read and to keep correct. Google recommends it. Microdata still works, and this validator lists the microdata types it sees without checking them in depth.

Can schema markup hurt my site?

Wrong markup is mostly ignored. The risk is markup that describes things the page does not show, such as reviews or prices that are not there. Google treats that as spam and can stop showing rich results for the site.

What does sameAs do?

sameAs lists the other places where your brand has a profile, such as its pages on social networks, business directories and encyclopedias. It tells a machine that all of them are the same entity. For a brand that wants AI to recognise it, this is the most useful property after the name.

How often should I validate my schema markup?

After every change to a template, a plugin or a theme, and after a redesign. Markup breaks quietly: the page looks the same and one missing comma removes the whole block.

Clean schema markup tells a machine who you are. Whether the assistants name your brand when someone asks is a separate question, and the AI Visibility check answers it.

Check your AI visibility