How to use the Schema Markup Validator

This free schema markup validator checks your structured data before search engines and AI systems read it. Paste a block of JSON-LD, or the HTML source of a page that contains JSON-LD script tags, and it reports syntax errors, missing properties and badly formatted values for every item it finds. It runs entirely in your browser, so nothing you paste is sent to a server or stored.

  1. Get the code. Copy the JSON-LD from your SEO plugin, theme or generator. For a live page, view the page source and copy everything. The validator finds each <script type="application/ld+json"> block in the HTML on its own.
  2. Paste and run the check. A single object, an array of objects and a document that uses @graph all work. If the JSON cannot be parsed, you see the error with its line and column.
  3. Review the list of types. The validator lists every item it found, including items inside @graph and items nested inside other items, with the type of each one.
  4. Work through each item. Every item gets its own errors, warnings and passes. Fix the markup at its source, paste the new version and check again.
  5. Confirm with Google and schema.org. Test the page in Google’s Rich Results Test and in the Schema.org validator at validator.schema.org as well. Each tool answers a slightly different question, as explained below.

Tip: If a page’s markup is added by JavaScript after the page loads, it may not appear when you view the source. Copy the rendered HTML from your browser’s developer tools instead, or test the live URL in Google’s Rich Results Test, which renders the page first.

What the validator checks

JSON syntax, with line and column

JSON-LD is JSON, and JSON is strict. One trailing comma, one curly quote or one missing bracket makes the whole block unreadable, and a parser that cannot read a block ignores all of it. The validator reports the error with its line and column, so you do not have to hunt for it. When you paste a full page, each script block is parsed separately, so one broken block does not hide problems in the others.

Every item, including @graph and nested items

Structured data is rarely one flat object. A product page may hold a Product with an Offer, an AggregateRating and several Reviews inside it, and an SEO plugin often outputs a @graph holding an Organization, a WebSite, a WebPage, a BreadcrumbList and an Article together. The validator walks the whole tree and lists every type it finds. That list alone shows what a parser actually sees, which makes duplicates and missing items easy to spot.

For common types, the validator checks whether the properties that matter are present. It covers Organization, LocalBusiness, WebSite, WebPage, Person, Article, BlogPosting and NewsArticle, Product with its Offer, AggregateRating and Review, SoftwareApplication, FAQPage, HowTo, BreadcrumbList, Event and VideoObject. Other types are still found and listed, but the property checks are written for these.

Value formats

A property can be present and still unusable, so the validator also checks the formats that most often go wrong:

  • Context. A missing @context, or one that does not point to schema.org, is flagged.
  • Dates. Properties such as datePublished, startDate and uploadDate should use ISO 8601, such as 2026-03-14 or 2026-03-14T09:30:00+01:00, not “March 14, 2026”.
  • URLs. Properties such as url, logo, image and item should be absolute, such as https://example.com/logo.png, not /logo.png.
  • Price and currency. A price should be a plain number such as 29.00, and priceCurrency a three-letter ISO 4217 code such as USD.
  • Ratings. A ratingValue should sit inside the stated bounds. With bestRating left out, the usual default is 5, so a rating of 9 is out of range.

Errors, warnings and passes

Results are grouped by item. An error means something is broken or a required property is missing, so the item may be ignored or ruled out of a feature. A warning means the markup works but is incomplete or looks suspicious, such as a missing recommended property. A pass confirms a check succeeded. Aim for zero errors, and treat warnings as a to-do list for information you actually have.

What a schema markup validator can and cannot tell you

Every schema markup validator answers a narrower question than people hope. This one tells you whether your JSON-LD is well formed, whether it has the properties common consumers expect and whether the values are in usable formats. It does not tell you how a search engine will treat the page.

  • It is not Google’s Rich Results Test. Google’s tool renders the page with Google’s own systems and reports which Google features the markup is eligible for. Only Google can answer that.
  • It does not guarantee rich results. Valid markup makes a page eligible to be considered. Whether a rich result appears depends on the search engine, the query and the page.
  • It does not cover the whole vocabulary. The Schema.org validator checks every type and property in schema.org and also reads microdata and RDFa.
  • It cannot see your page. It only knows what you paste, so it cannot tell whether the markup matches the visible content or whether an image URL loads.

Eligibility also changes. In 2023 Google limited FAQ rich results to well-known, authoritative government and health websites, and it stopped showing HowTo rich results. That markup is still valid and still helps machines understand the page, but most sites should not expect a visual result from it. Check Google Search Central for the current list of features before planning work around one.

A note on names: the tool at validator.schema.org is officially called the Schema Markup Validator too. This aeoora tool is a separate, independent checker, and the two work well together.

What structured data is and why it matters for AI answers

Structured data is a machine-readable description of what a page is about. Instead of leaving a parser to work out from the layout that a page sells a product for 49 euros, has a 4.6 rating and is in stock, you state those facts in a standard format with standard names.

The names come from schema.org, a shared vocabulary launched by Google, Microsoft and Yahoo in 2011, with Yandex joining later that year. It defines types, such as Organization, Product and Event, and the properties each can have, such as name, price and startDate. JSON-LD is the most common way to write it: a block of JSON in a script tag, separate from the visible HTML.

Entities and knowledge graphs

Search engines use structured data to understand pages and to power enhanced results such as product details, event listings and breadcrumbs. They also organize facts around entities: specific companies, people, products and places. A knowledge graph is a database of those entities and how they relate. Organization markup with your official name, logo, URL and sameAs links to your profiles states facts about your brand in a form these systems can match against other sources, and Person markup does the same for authors. Consistency matters: when every page describes your company with the same name and URL, a machine can connect them with confidence.

Why it matters for AI answers

AI assistants such as ChatGPT, Perplexity, Gemini, Claude and Google’s AI Overviews answer from what they learned in training and from pages they retrieve at the time of the question. No AI company publishes exactly how structured data feeds into its answers, so be wary of promises of a direct effect on citations. What is clear is that many answer engines lean on search indexes and knowledge graphs that do use structured data, and that explicit, consistent facts are harder to misread than facts buried in a layout. Pair clean markup with clear writing, which you can test with the AI Answer Readiness Checker.

Schema.org itself has no required properties. Every property is optional in the vocabulary. Requirements come from the consumers of the markup: Google, for example, documents for each rich result type which properties are required for eligibility and which are recommended. The validator follows the same idea:

  • Required properties are the ones a type cannot do without for its common uses, such as the address of a LocalBusiness or the startDate of an Event. Missing ones are errors.
  • Recommended properties add useful detail, such as an image or a dateModified. Missing ones are warnings.
  • Nested requirements apply to items inside items. An Offer needs a price and a currency, and an AggregateRating needs a ratingValue and a count.

Never add a property just to clear a warning. Google’s guidelines say structured data must represent the content visitors see, and misleading markup can lead to a manual action. If a product has no reviews yet, leave review and aggregateRating out.

The most common schema errors and how to fix them

Most structured data problems come from a short list of causes.

Invalid JSON

The usual causes are a trailing comma after the last property, single quotes instead of double quotes, curly quotes pasted from a document editor, an unescaped double quote inside a text value, a comment (JSON does not allow them) or a missing bracket. Go to the reported line and column. If it points at the start of a line, the real problem is often at the end of the line before.

Missing or incorrect context

Every top-level block needs "@context": "https://schema.org", and copying a fragment out of a larger block often leaves it behind. Typos and values such as schema.org without a protocol can stop a parser from mapping your terms. The older http://schema.org form is still widely accepted, but use https in new markup.

Misspelled types and properties

Schema.org names are case-sensitive: BlogPosting is a type, Blogposting is not, and datepublished is not datePublished. If the type list shows a spelling you do not recognize, fix it at the source. The Schema.org validator is the best second check, because it knows every term.

Dates, URLs, prices and ratings

  • Dates: use 2026-03-14T19:00:00-05:00, not a display format such as “14/03/2026”. Durations, such as a video’s length, use the form PT4M12S.
  • URLs: replace relative paths such as /images/cover.jpg with full URLs, and check that they load. The validator can tell a URL is absolute, not that the file exists.
  • Prices: write "price": "1299.00" with "priceCurrency": "USD", not "$1,299". Use a period for decimals and no thousands separator.
  • Ratings: if you rate out of 10, say so with "bestRating": "10", and give an AggregateRating a ratingCount or reviewCount.

Duplicate or conflicting markup

A theme, an SEO plugin and a page builder often each output their own Organization or Article. The type list makes this obvious: two Organizations with different logos, or an Article and a BlogPosting for the same post. Pick one source per entity, switch off the others and connect what remains with @id references.

Markup that does not match the page

A price that differs from the one on the page, an author missing from the byline or reviews visitors cannot see do not break the JSON, so no syntax check catches them. They are a policy problem. Compare the markup with the rendered page, especially after you change prices, dates or authors.

Type-by-type notes

These notes cover what each checked type is for and what to include. For the exact requirements of a particular Google rich result, follow Google’s current documentation, because it changes.

Organization and LocalBusiness

Organization describes your company: name, url, logo, sameAs links to official profiles and contactPoint. Define it once, on the home page or About page, and reference it elsewhere. LocalBusiness, or a subtype such as Restaurant or Dentist, is for places customers visit. Name and address are the core, and telephone, opening hours, geo coordinates and image make it useful. The Organization Schema Generator builds a clean starting block.

WebSite, WebPage and Person

WebSite usually appears once, on the home page, with a name and url, and Google uses it among other signals to choose the site name shown in results. Google has retired the sitelinks search box, so a SearchAction no longer produces that feature. WebPage describes an individual page, and SEO plugins often use it in a @graph to tie the page to its breadcrumb and site. Person describes authors and founders: give a name, a url to a profile page, sameAs links, a jobTitle and worksFor pointing to your Organization.

Article, BlogPosting and NewsArticle

Use BlogPosting for blog posts, NewsArticle for news and Article for other editorial content. Google recommends headline, image, datePublished, dateModified and author. Keep the headline the same as the visible title, change dateModified only for real updates and make the author match the byline. The Article Schema Generator produces a complete block.

Product, Offer, AggregateRating and Review

A Product needs a name and, to be useful in search, at least one of an Offer, an AggregateRating or a Review. The Offer carries price, priceCurrency and availability, using values such as https://schema.org/InStock. An AggregateRating needs a ratingValue plus a ratingCount or reviewCount, and each Review needs an author and a reviewRating. Add image, description, brand and identifiers such as sku or gtin when you have them.

SoftwareApplication

Use SoftwareApplication, or WebApplication or MobileApplication, for software. Include the name, an Offer with a price (use 0 for a free app) and a rating or review if you have genuine ones. The applicationCategory and operatingSystem properties say what kind of software it is and where it runs.

FAQPage and HowTo

FAQPage holds a mainEntity list of Question items, each with a name and an acceptedAnswer containing an Answer with text. Use it only for questions the site answers itself, not user forum threads. HowTo describes a process with a name and a list of HowToStep items, each with text. Google now limits FAQ rich results and no longer shows HowTo rich results, so the value of both is clarity, not a snippet. The FAQ Schema Generator builds FAQPage markup from your questions.

A BreadcrumbList has an itemListElement array of ListItem entries, each with a position starting at 1, a name and an item URL. The last item, the current page, can leave the URL out. Match the trail visitors see and keep positions in order without gaps.

Event and VideoObject

An Event needs a name, a startDate with a time zone offset and a location: a Place with an address, or a VirtualLocation with a url for online events. Add endDate, eventStatus, eventAttendanceMode, image, offers and organizer, and update eventStatus if plans change. A VideoObject needs a name, a thumbnailUrl and an uploadDate, and a description, an ISO 8601 duration and a contentUrl or embedUrl help search engines understand the video itself.

JSON-LD vs microdata

There are three common ways to add schema.org data to a page:

  • JSON-LD is a JSON block inside a <script type="application/ld+json"> tag, separate from the visible HTML.
  • Microdata puts attributes such as itemscope, itemtype and itemprop on the visible elements.
  • RDFa works much like microdata, with attributes such as typeof and property.

Google supports all three and recommends JSON-LD where possible. It is easier to write and maintain because it is not woven through your templates: a designer can change the layout without breaking the markup, and nested items connect cleanly with @graph and @id. Microdata’s one advantage is that, being attached to visible elements, it is less likely to drift from the content.

This validator reads JSON-LD only. For microdata or RDFa, use the Schema.org validator or Google’s Rich Results Test, which read all three. On a new site, choose JSON-LD, and avoid describing the same entity in two formats, which tends to create duplicates.

Testing workflow: where a schema markup validator fits

A schema markup validator is most useful as one step in a routine that catches problems while they are cheap to fix.

  1. Validate while drafting. Paste JSON-LD here as you write it or configure a plugin, and fix syntax errors and missing required properties before anything goes live.
  2. Check the whole page. Paste the full HTML of a staging or live page to see every block, including those added by your theme and plugins.
  3. Run Google’s Rich Results Test. See which items Google detects and which features they are eligible for. A valid FAQ or HowTo item may show as not eligible because of Google’s policy changes, not an error.
  4. Run the Schema.org validator. Catch unknown types and properties, and check any microdata or RDFa.
  5. Compare the markup with the page. Prices, dates, names, ratings and answers should match what visitors see.
  6. Watch Search Console and re-test after changes. Check the rich result reports after Google recrawls, and test again after theme updates, plugin changes and migrations, the usual causes of broken markup.

On a large site, test one representative page per template. If the product template is right, most product pages will be too, apart from data gaps such as a missing price or image.

Frequently asked questions

Is this the same as Google’s Rich Results Test?

No. This validator checks syntax, properties and formats in your browser. Google’s Rich Results Test shows how Google reads a page and which Google features it is eligible for. Use this tool to find and fix problems quickly, then confirm with Google’s test.

Does valid markup guarantee rich results?

No. Valid markup makes a page eligible to be considered, nothing more, and search engines change their policies. FAQ rich results in Google, for example, are now limited mostly to authoritative government and health sites.

Can I check a URL instead of pasting code?

The validator works on what you paste and does not fetch pages. View the page source, copy everything and paste it in. If the markup is added by JavaScript, copy the rendered HTML from your browser’s developer tools, or test the URL in Google’s Rich Results Test.

Does it check microdata or RDFa?

No, it reads JSON-LD only. For microdata and RDFa, use the Schema.org validator at validator.schema.org or Google’s Rich Results Test.

Why do other validators show different results?

They check different rules. Google’s test checks the requirements for Google features, the Schema.org validator checks the full vocabulary, and this validator checks common properties and formats. A property can be optional in one and expected in another.

Is the code I paste sent anywhere?

No. The checks run in your browser, and the code is not uploaded or stored. It is gone when you close the page.

Does structured data help with ChatGPT and other AI assistants?

It can help indirectly. AI companies do not publish how they use it, but many answer engines draw on search indexes and knowledge graphs that read it. Treat markup as one part of being clearly understood, not as a switch that produces citations.

Which schema types should my site use?

Start with Organization or LocalBusiness and WebSite on the home page, plus BreadcrumbList across the site. Then match each template: Article or BlogPosting for posts, Product for products, Event for events and VideoObject for video pages. Only mark up what the page shows.

What is @graph, and is it valid?

Yes, @graph is valid JSON-LD. It lets one script hold several items that share a @context and refer to each other by @id, such as a WebPage whose publisher is the Organization in the same graph. The validator checks every item inside it.

Structured data works best alongside the rest of the page. These free tools cover the related pieces.

Browse the full set on the free tools hub.

Go further with aeoora

Valid markup helps machines read your pages correctly. It does not tell you whether AI assistants actually mention or recommend you. The free AI visibility report checks that for your brand across 25 buyer questions and emails you a PDF, so you can see where you appear and where competitors appear instead.

When you want to follow it over time, start a 2-day trial of aeoora Cloud. It tracks your mentions and citations across ChatGPT, Perplexity, Gemini, Claude and Google AI Overviews, and its Site Audit flags structured-data problems across your whole site rather than one page at a time. You can compare plans on the pricing page.