Create JSON-LD schema output in seconds—upload a supported input and generate a JSONLDSCHEMA file directly in your browser.
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": "Article"
}
</script>Article that's headline, image, author, datePublished; for Product it's name, image, plus at least one of review / aggregateRating / offers; for LocalBusiness it's name, address, telephone. Add optional fields (description, url, sameAs, dateModified) to enrich the markup — they don't trigger errors if omitted but they help Google and AI crawlers disambiguate the entity.<script type="application/ld+json">…</script> block appears with the correct "@context": "https://schema.org" header, the right "@type", and your values escaped properly. The block updates as you type — no "build" button needed.<head> (the convention) or anywhere in <body> — Google parses both. Validate at the Rich Results Test before shipping. Everything runs client-side: no upload, no account, no rate cap.Structured data is how you tell search engines and LLMs what your page actually is — a recipe, a product listing, a local business, a job posting, an event — separate from the prose a human reads. Google, Bing, Yandex, Pinterest, and increasingly the AI answer engines (ChatGPT search, Perplexity, Claude's web tool) all parse schema.org markup to populate knowledge panels, sidebars, AI Overviews, and entity cards. JSON-LD is the format every major search engine prefers because it lives in <head> as a self-contained block instead of being woven through your HTML, which means a template change can't accidentally break it.
sameAs links to Wikidata, LinkedIn, and verified social profiles.Pair this with the JSON Formatter to pretty-print or validate the generated output, the HTML Formatter when embedding the script block into a page template, and the HTML Minifier before shipping production HTML.
Schema.org defines 823 types as of version 30.0 (March 2026), but only ~30 produce visible rich results in Google Search. Here's the current state for the most-requested ones:
| Schema type | Rich-result status (2026) | What it enables |
|---|---|---|
Article / NewsArticle / BlogPosting |
Supported | Top Stories carousel, headline image in SERP, Google News inclusion |
BreadcrumbList |
Supported | Replaces the URL line in SERP with Home > Category > Page |
Event |
Supported | Event card with date/venue, eligible for Google's events experience |
FAQPage |
Deprecated — rich results removed May 7, 2026; appearance + Rich Results Test support removed June 2026; Search Console API removed August 2026 | Previously: dropdown FAQ snippet under blue link (still works for some .gov/.edu/health domains as a manual exception) |
HowTo |
Deprecated — removed from desktop September 13, 2023 and from mobile in the same Aug-2023 announcement; documentation deleted by Google in 2026 | Previously: numbered step carousel in SERP |
JobPosting |
Supported | Google for Jobs box at top of SERP |
LocalBusiness |
Supported | Knowledge panel, Maps integration, hours/address card |
Product |
Supported | Price, rating, availability badge under blue link; merchant listing experience |
Recipe |
Supported | Recipe rich card with image, rating, cook time; eligible for Google Discover food module |
Review / AggregateRating |
Supported (only embedded in Product, LocalBusiness, Book, etc.) | Star rating under the blue link |
VideoObject |
Supported | Video thumbnail, key moments, live-badge in SERP and Discover |
Organization / Person |
Supported (knowledge-panel only, not a SERP rich result) | Feeds the knowledge panel sidebar; sameAs consolidates entity signals |
WebSite with SearchAction |
Supported | Sitelinks search box under your homepage in SERP |
The deprecation of HowTo and FAQPage does not mean removing the markup — Google's own guidance is that unused structured data causes no harm. Many publishers leave the FAQPage markup in place because Bing, DuckDuckGo, and AI answer engines still parse it, and the .gov / health-domain exception in Google means a small set of sites still see the card.
| Property | JSON-LD | Microdata | RDFa |
|---|---|---|---|
| Where it lives | Standalone <script> block, typically in <head> |
Inline HTML attributes (itemscope, itemtype, itemprop) on existing elements |
Inline HTML attributes (vocab, typeof, property) |
| Coupling to markup | Decoupled — change visible HTML without touching schema | Tightly coupled — refactoring template breaks schema | Tightly coupled |
| Google support | Recommended | Supported | Supported |
| JavaScript-rendered pages | Works (script tag survives client-side hydration) | Fragile — only what Googlebot sees in rendered HTML counts | Fragile |
| Adoption (Web Almanac 2024) | 41% of pages (up from 34% in 2022, fastest-growing) | ~22% | ~66% (largely legacy CMS like Drupal/Open Graph dialect) |
| Maintainability | High — one block per page, easy to template | Low — schema scattered across the DOM | Low |
All three formats are equally parsed by Google. The reason JSON-LD has been the recommended default since 2015 and the dominant format in new content is purely operational: a designer changing the visible HTML cannot accidentally invalidate it.
No, leave it. Google's official guidance is that "structured data that's not being used does not cause problems for Search, but also has no visible effects." HowTo rich results were removed from mobile in early August 2023 and from desktop on September 13, 2023; Google deleted the HowTo documentation entirely in 2026. But Bing still renders HowTo in some result types, and AI assistants (Perplexity, Bing Copilot, Claude's web tool) parse HowTo blocks as a structured signal about the page's intent. Strip the markup only if you're cleaning up bytes — there's no SEO downside to leaving it.
Yes. Google deprecated FAQ rich results for everyone except authoritative government and health domains in August 2023. As of May 7, 2026, FAQ rich results stopped appearing in Search entirely. The FAQ appearance and rich result report in Search Console will be removed in June 2026, and the Search Console API will drop FAQ support in August 2026. The schema.org FAQPage type itself is not deprecated — it stays in the vocabulary — so other consumers (Bing, AI engines, in-house search) can still read it. Just don't expect the dropdown snippet under your blue link.
Officially Google says all three are "equally fine" and that they parse them identically — confirmed in Google Search Central documentation. In practice JSON-LD wins because it's decoupled from the visible HTML: a designer changing the template can't break it, JavaScript frameworks can inject it cleanly into <head>, and the entire block is in one place for review. The Web Almanac 2024 found JSON-LD on 41% of pages (up from 34% in 2022) and growing fastest. RDFa still appears on 66% of pages, but almost entirely on legacy Drupal sites where it was templated in years ago.
<script> block go — <head> or <body>?Either works for Google. The <head> is the convention because it groups all metadata in one place and is the first thing parsers see, but Google's Structured Data documentation explicitly states the block "can be placed anywhere on the page." For JavaScript-rendered single-page apps, injecting the block via next/script or a <Helmet> component into <head> is the cleanest pattern — for static pages, server-rendering it into <head> ensures Googlebot sees it on the first crawl without waiting for JS execution.
Google publishes a per-type list. For example, Article requires headline, image, author, datePublished; Product requires name, image, and one of offers/review/aggregateRating; Recipe requires name, image, plus at least three of nutrition, aggregateRating, cookTime, recipeIngredient. Missing a required field shows as an error in the Rich Results Test and disqualifies the page from the rich result. Missing a recommended field shows as a warning — the rich result still triggers, but enrichments like rating stars or cook-time badges may not appear. The generator marks required fields and only lets you copy a valid block.
Use Google's Rich Results Test for rich-result eligibility (Google-specific — tells you if Product, Article, Recipe, etc. will trigger the SERP feature) and the Schema Markup Validator for generic schema.org compliance (parses any vocabulary type and reports syntax errors without Google's eligibility opinion). Run both — Rich Results Test catches Google-specific requirements (e.g., priceCurrency must be ISO 4217), and the Schema Markup Validator catches vocabulary mistakes (@type: "Persoon" typos, wrong namespace). Both accept either a URL or a raw paste.
The current schema.org vocabulary is version 30.0, released March 19, 2026, with 823 types, 1,529 properties, 96 enumerations, and 535 enumeration members. Schema.org ships updates roughly quarterly (29.0 in March 2025, 29.4 in December 2025, 30.0 in March 2026). Google and the other consumers track the latest stable version automatically — there's no per-page version pin. The vocabulary itself is backwards-compatible across major versions, so older blocks continue to parse correctly.
Yes — and this is the recommended pattern for any page that has more than one entity to describe. Use a @graph array inside a single <script> tag, listing each top-level entity as an object. Google's documentation explicitly supports this and recommends one combined script over multiple separate scripts, since it reduces parser overhead and lets you cross-reference entities with @id (e.g., the Article's publisher can point to the Organization object's @id in the same graph). Most generators produce one type at a time; if you need multi-type output, generate each separately and merge them into a single @graph wrapper by hand.
No. The generator runs entirely in your browser — your page titles, addresses, prices, and any other field values never leave the tab. No account, no log, no rate limit, no watermark on the output.