In short. JSON-LD is a JSON-based format for Linked Data that websites can embed in a script element to describe page content.

Definition: JSON-LD is a JSON-based format for expressing Linked Data: machine-readable data connected across documents and websites. The W3C specification defines its syntax, including identifiers and a context that gives data fields a shared meaning.

Key facts

The W3C recommendation for JSON-LD 1.1 is dated 16 July 2020.
As of 17 October 2026, Google's structured data introduction recommends JSON-LD for website structured data and also supports the HTML-based formats Microdata and RDFa.
Google's structured data policies require markup to represent the page's content accurately and describe content visible to readers.
Google's AI features documentation states that AI Overviews and AI Mode require no special schema.org structured data.

What does JSON-LD look like on a page?

The W3C specification's HTML examples embed JSON-LD in a script element with the type application/ld+json. The following block reproduces the person data from its example describing Dave Longley:

<script type="application/ld+json">
{
  "@context": "http://schema.org/",
  "@id": "https://digitalbazaar.com/author/dlongley/",
  "@type": "Person",
  "name": "Dave Longley"
}
</script>

The @context supplies the vocabulary used to interpret short names. Here, Person identifies the type of thing described, and name supplies its name. The @id identifies the described person with a web identifier. These roles come from the specification's definitions of context, type and node identifiers.

The block is an example of the format. Its properties describe a person; a website implementation needs properties appropriate to the content it represents.

How does JSON-LD relate to schema.org?

Structured data is a standardized way to describe and classify page content. In the example above, JSON-LD supplies the syntax, while schema.org supplies the vocabulary used for the person description.

Google's introduction to structured data explains that most Search structured data uses schema.org vocabulary. Google uses this information to understand page content, including information about people or companies.

For a specific Google search feature, the same documentation directs implementers to Search Central's required and recommended properties. A rich result is a search result with an enhanced display. Eligibility for that display requires all properties marked as required for the relevant object.

How should a website check its markup?

Google's structured data introduction recommends the Rich Results Test during development and rich result status reports after deployment. The latter help identify markup that breaks because of template or serving changes.

Google's general policies also require a content review. If JSON-LD describes a performer, the visible HTML must describe that same performer. Required properties must be complete, and pages containing the markup must remain accessible to Googlebot. The policies explain that automated tools cannot easily test the quality guidelines, so a technical test needs this editorial check alongside it.

FAQ

Where can I put JSON-LD in an HTML page?

Google's format documentation supports a script element in the page's head or body. The markup can sit separately from the visible text it describes.

Can Google read JSON-LD added with JavaScript?

Yes. The same format documentation states that Google can read JSON-LD dynamically injected by JavaScript or embedded content management system widgets.

Do I need special JSON-LD for Google's AI search features?

Google's AI features guidance specifies no special schema.org markup for AI Overviews or AI Mode. It asks you to keep structured data consistent with visible page text.

Sources

Check how a page reads to answer engines. Open the GEO score analyzer