Structured data checker: read the JSON-LD of a page

How to use the tool

Paste the full address of a page and click “Check”. The tool downloads the page the way a browser would, follows redirections to the destination and keeps only the structured data blocks of the HTML: the server does not send the whole page back to you. If the page is protected, or if you want to test markup before publishing it, open “Paste JSON-LD or HTML instead”.

Each block found gets its own card: syntax status, structure alerts, then a paragraph per type with the missing properties, required ones first. The folded property tree shows what the tool actually read, which helps to spot a badly nested value. The tool reads up to 60 blocks per page and counts one request per address.

The three formats the tool reads

Google accepts three formats: JSON-LD, microdata and RDFa. It treats them as equivalent as long as the markup is valid and follows the documentation of the targeted feature, and it recommends JSON-LD: a <script> block separate from the visible text, which suits nested data and is the easiest to maintain at scale.¹ The tool handles the three the same way: it converts microdata (itemscope, itemtype, itemprop) and RDFa (typeof, property) into objects shaped like JSON-LD, then applies the same checks.

For JSON-LD it also reports what belongs to the structure of the block: a missing @context or one that is not schema.org, an object with neither @type nor @graph, an @id defined twice, a reference to an @id the block does not define. Valid references between objects of the same @graph are resolved before the comparison: an author returned by {"@id": "#me"} counts as present, with the properties of the object it points to.

Finding the line of a syntax error

A trailing comma after the last property, a typographic quote, a comment: JSON is strict, and a single faulty character invalidates the whole block. When the browser reports the position of the error, the tool converts it into a line and column number inside the block and shows that line. The browser’s original message stays visible underneath, in English. A block with an error is not compared with Google’s properties: fix the syntax first.

What Google lists for each type

For the types that open a rich result, Google’s documentation separates required properties, without which the page is not eligible, from recommended properties, which enrich the display. The table below reproduces what Google’s pages say for each type; the tool compares your markup with those lists, nested paths included.²

Three remarks on this table. Article and Organization have no required property: Google asks you to add the ones that match your content. In a breadcrumb, item may be omitted on the last element, in which case Google uses the page address. For a Product, any one of review, aggregateRating or offers is enough.⁴⁷ Subtypes of LocalBusiness, Organization and Article are recognised for a common list (Restaurant, Store, NewsArticle, BlogPosting and a few others); the full schema.org vocabulary has many more.

The limits of the tool

Following the property list does not guarantee a display: Google also requires the markup to describe content visible on the page, the information to be up to date and reviews to be genuine user reviews. Markup that breaks these rules risks a manual action, which removes eligibility for rich results without changing the ranking in web search.² The tool reads code, it does not compare the markup with the displayed content.

Google’s Rich Results Test remains the reference to know whether a page is eligible, and the schema.org validator checks the markup against the full vocabulary.¹¹² Use the tool to sweep a page quickly, read the tree and spot the gaps, then confirm with those two services.

Another limit: the tool reads the HTML as the server sends it, without running JavaScript. Markup injected by a script after loading does not show up; copy the code from your browser’s inspector and paste it instead. Finally, FAQ is no longer in Google’s search features gallery:¹³ a FAQPage block is read and its tree displayed, but it has no property table.

After the check

To write clean markup rather than repair broken markup, the structured data generator produces the JSON-LD for LocalBusiness, Article, breadcrumbs and FAQ with the fields Google expects. The article on structured data and rich snippets explains which types are worth adding. And when a whole site has to be reworked, type by type and template by template, that is the job of an SEO consultant.

Frequently asked questions