What Is JSON-LD?
Updated 4 November 2026
Jump to section
JSON-LD is a way of writing structured data as a self-contained block of JSON inside your page's code, and it is the format Google recommends for schema markup. The name stands for JSON for Linking Data. Instead of weaving labels through your visible HTML, JSON-LD sits in one script tag. It describes what the page is about, such as a local business, a product, or an article.
What does JSON-LD look like?
JSON-LD looks like a small block of code in a <script> tag, usually placed in the page's <head>. It holds facts about the page as simple name-and-value pairs.
Here is an illustrative example for an article, not a live snippet:
{
"@context": "https://schema.org",
"@type": "Article",
"headline": "What Is JSON-LD?",
"author": { "@type": "Organization", "name": "Storming Solutions" },
"datePublished": "2026-08-28"
}
Two keys carry the meaning. @context names the vocabulary, almost always schema.org. @type says what kind of thing the page describes, here an Article.
Google's documentation, updated December 2025, describes JSON-LD as a script embedded in the head or body of a page. A visitor never sees it, but a search engine reads it directly.
Is JSON-LD the same as schema.org?
No, JSON-LD and schema.org are two different things that work together. JSON-LD is the format, the way you write the data. Schema.org is the vocabulary, the shared dictionary of types like Product, Article, and LocalBusiness.
The split is worth holding on to. In the words of the W3C standard, a Recommendation from July 2020, JSON-LD is a lightweight syntax to serialize Linked Data in JSON. Schema.org supplies the words; JSON-LD is the grammar that arranges them.
So you write schema.org types using the JSON-LD format. The two names describe different layers of the same markup, not rival choices.
Why does Google recommend JSON-LD?
Google recommends JSON-LD because it is the easiest structured-data format to add and maintain without breaking. It lives in one block, separate from your visible content.
The same documentation lists three supported formats: JSON-LD, Microdata, and RDFa. Microdata and RDFa thread their labels through the visible HTML, so a design change can easily break them. JSON-LD stays in its own script, which is why Google calls it the easiest to implement and maintain at scale.
| Format | Where it lives | Google's stance |
|---|---|---|
| JSON-LD | a single script block | recommended |
| Microdata | attributes inside visible HTML | supported |
| RDFa | attributes inside visible HTML | supported |
Source: Google Search Central, intro to structured data, updated December 2025.
In our experience, the markup and the visible page drift apart over time, so a block that was right at launch can quietly go stale. A redesign is a good moment to re-test it, since a new template can drop or change the markup.
Valid markup is also no promise of a rich result. Google removed FAQ rich results from Search on 7 May 2026, yet FAQPage markup is still valid.
Markup also comes second to the page itself. Content quality earns citations, while schema only helps display them, and no markup gets a page into an AI Overview.
Frequently asked questions
Does JSON-LD help my Google ranking?
Not directly. JSON-LD is a format for structured data, and structured data makes a page eligible for rich results rather than lifting its rank on its own. The gain is indirect but real: a richer, clearer listing can earn more clicks, and machine-readable facts are easier for search engines and AI answer engines to quote. The value is in being understood, not in a ranking boost.
Where do I put JSON-LD on my page?
Inside a <script type="application/ld+json"> tag, usually in the page's <head>. Google can read it in the <head> or the <body>, so either works, but the head is the common choice. Because it sits in its own block and never appears to visitors, it does not affect how the page looks. Many content platforms and SEO plugins insert it for you automatically.
Is JSON-LD better than Microdata?
Usually, yes. Google recommends JSON-LD because it lives in one script block, separate from your content, so it is simpler to add and less likely to break when the design changes. Microdata still works and Google supports it, but its labels are mixed into the visible HTML, which makes it fiddly to maintain. New markup is usually easier to build in JSON-LD.
Do AI answer engines read JSON-LD?
It can help, but it is not required. Google's AI guidance puts it plainly: "Structured data isn't required for generative AI search", though it still recommends markup for rich-result eligibility. JSON-LD gives any machine a clean, labeled summary of a page, which makes your facts easier to read accurately. It is one signal among many and never replaces clear on-page content. Still, well-structured data supports how AI engines choose what to cite, because accurate labels are easier to trust.
Do I need to write JSON-LD by hand?
Rarely. Many website platforms, and SEO plugins on systems like WordPress, generate JSON-LD from the details you enter, so you fill in a form rather than write code. Hand-writing it is only needed for custom types or unusual setups. Whether it is generated or written, the rule is the same: the markup must match what visitors actually see on the page.
Get the labels machines read right
Storming Solutions runs SEO, AEO, and GEO for Malaysian businesses from Kuala Lumpur. We would rather mark up only what genuinely appears on the page than chase every schema type, because accurate structured data is what earns trust from an answer engine.
Not sure what structured data your site carries, or whether a redesign stripped it out? Ask us through the contact page, or read what schema markup is next. Our SEO, AEO, and GEO service makes sure the labels machines read match the page people see.