DefinedTerm · Glossary
What is JSON-LD
JSON-LD (JavaScript Object Notation for Linked Data) is a W3C standard serialisation format for Linked Data that uses plain JSON with a special @context key to attach meaning to data fields. In web SEO, JSON-LD is the recommended method for embedding Schema.org structured data into HTML pages. Because the block lives inside a <script> tag and never touches the visible markup, it can be injected server-side or at build time without altering design. Google, Bing, and AI assistants parse JSON-LD to extract verified entity facts — business name, address, reviews, products — and use them in rich results and knowledge-graph entries.
Full definition
JSON-LD stands for JavaScript Object Notation for Linked Data. It is a serialisation format specified by the W3C that allows developers to express Linked Data — data whose fields carry explicit, machine-interpretable meanings — using standard JSON syntax.
The mechanism is simple: a special @context key points to a vocabulary (most commonly https://schema.org) that defines what every property name means. A @type key identifies the class of the described entity (LocalBusiness, FAQPage, Product, and so on). All other keys are properties drawn from that vocabulary.
On web pages, the block is placed inside a <script type="application/ld+json"> element, typically in the <head>. The browser ignores it; crawlers parse it. Because the block is entirely separate from the visible DOM, it can be generated by server-side code, injected by a CMS plugin, or produced at build time without touching the HTML design.
Why it matters in 2026
Google officially recommends JSON-LD over Microdata and RDFa for new implementations and states that its parser handles JSON-LD correctly even when rendered by JavaScript, provided Googlebot can execute the page. This matters for Next.js and other JavaScript-framework sites where the schema can be injected into the <head> component server-side, guaranteeing crawlability without depending on client-side hydration.
Beyond classic rich results, JSON-LD structured data feeds the entity layer consumed by AI-generated answer systems. Google AI Overviews, Bing Copilot, and third-party AI assistants that index the web use the entity graph — substantially populated by JSON-LD — to retrieve factual attributes when composing answers. A business with complete, error-free JSON-LD is more likely to have its facts quoted accurately and attributed directly in AI responses.
How it works
A crawler visits the page and extracts the <script type="application/ld+json"> block. The JSON is validated against the declared @context vocabulary. Valid properties are mapped to Knowledge Graph nodes and edges. Errors — unknown properties, wrong value types, missing required fields — cause the block to be partially or fully rejected. Google Search Console reports these under the Enhancements section.
Nested objects such as address, openingHoursSpecification, and geo must themselves carry a @type to be interpreted correctly. Arrays are valid for repeating properties such as openingHoursSpecification (multiple day ranges) or sameAs (multiple external profiles).
Difference from Microdata and RDFa
| Format | Placement | Syntax overhead | JavaScript-rendered support | Google recommendation |
|---|---|---|---|---|
| JSON-LD | <script> in <head> or <body> | Minimal — pure JSON | Yes | Preferred |
| Microdata | Inline HTML attributes | High — modifies every element | Depends on crawler | Supported |
| RDFa | Inline HTML attributes | High — verbose namespace declarations | Depends on crawler | Supported |
Related terms
LocalBusiness schema, sameAs, E-E-A-T.
Fuentes
Términos relacionados
- localbusiness-schema
- sameas
- e-e-a-t