Structured Data Validator
Paste a URL or raw JSON-LD and get per-entity validation: syntax errors, missing required and recommended properties, and an honest note on what each schema type actually does for AI visibility.
How it works
- 1
Enter a URL or paste your markup
URL mode fetches the page and extracts every JSON-LD script block; paste mode accepts full HTML or a raw JSON-LD object and runs entirely in your browser.
- 2
We validate each entity
Blocks are parsed, @graph arrays flattened, and every entity is checked against per-type rules: required properties as errors, recommended ones as warnings, plus date, URL and rating sanity checks.
- 3
Fix errors first, then warnings
Syntax errors and missing required properties make markup unusable; recommended properties strengthen the entity signal. Each card tells you which AI surfaces the type actually feeds.
What structured data does for AI visibility — honestly
Structured data doesn't make AI engines recommend you. What it does is narrower and still valuable: it turns the entities on your pages — your brand, your authors, your products — into machine-readable statements that engines and LLM knowledge systems can resolve unambiguously. When an AI assistant assembles an answer about your category, entity resolution decides whether a mention of "Acme" connects to your Acme. Organization and Person markup with sameAs links are the strongest signals for that. Article markup carries authorship — the machine-readable half of E-E-A-T. Product markup feeds shopping surfaces. And FAQPage? Google deprecated FAQ rich results in May 2026, so the markup earns no rich result anymore — but the Q&A format itself remains highly quotable by AI assistants, and the markup stays valid for non-Google engines. We tell you this per type, in the results, instead of pretending every schema block is a growth lever.
The failure mode we see most: markup that's almost right. A missing @context, a required property omitted, a rating value outside its own declared bounds, a relative URL where crawlers need an absolute one. Parsers don't partially accept broken blocks — they skip them — so a single syntax error silently throws away everything the block declared. That's why this validator weights errors heavily: a page with three warnings is fine; a page with one syntax error has no structured data at all as far as machines are concerned.
How this validator works, and what it is not
We extract every <script type="application/ld+json"> block, parse it (reporting syntax errors with line and column), flatten @graph arrays into individual entities, and validate each entity against a rule table for the types that matter to AI visibility: required properties are errors, recommended properties are warnings, and we sanity-check dates (ISO-8601), URLs (absolute), and AggregateRating bounds. Microdata and RDFa are counted but not validated.
To be clear about what this is not: we are not Google's Rich Results Test and not the official schema.org validator. Use them as second opinions — the Rich Results Test checks only Google-supported rich result types against Google's requirements, and the schema.org validator checks pure vocabulary correctness. This tool sits between them: vocabulary plus documented per-type requirements, scored through an AI-visibility lens, with honest notes on which surfaces consume each type.
Best practices for 2026: one graph, linked entities
Put JSON-LD in the <head> where parsers reliably find it, and prefer one @graph block over scattered fragments. The current best practice for entity disambiguation is an @id-linked graph: your Organization gets @id: "https://yoursite.com/#organization", each author Person gets a stable @id, and every Article references them instead of restating them. Every page then reinforces one consistent set of entities — which is exactly what knowledge-graph construction rewards. Our Organization and Author schema generators emit these @ids for you.
Once your markup is clean, the real question is whether AI models connect those entities to your brand when customers ask. That's what our free AI Visibility Checker shows — live answers from grounded AI models, with your brand's presence scored.
Frequently asked questions
- How is this different from Google's Rich Results Test?
- Google's Rich Results Test only checks the schema types that can produce a Google rich result, and only against Google's requirements. This tool validates against schema.org vocabulary plus documented per-type requirements, focused on the types that matter for AI visibility — including ones Google no longer rewards, like FAQPage. Use both: they answer different questions.
- Which schema types does this validator check in depth?
- Organization (and subtypes like Corporation and LocalBusiness), Person, FAQPage, Article/BlogPosting/NewsArticle, Product, WebSite, WebPage, BreadcrumbList, HowTo, Review and AggregateRating — the roughly twelve types that carry entity, authorship, and product signals AI engines use. Other types get generic checks (syntax, @context, dates, URLs) and are flagged as outside the rule set.
- Does the paste mode upload my HTML anywhere?
- No. Paste mode runs entirely in your browser — the same validation code that checks fetched pages runs locally on your pasted content. Only URL mode makes a server request, and that request only fetches the page you name.
- What about microdata and RDFa?
- We detect and count them but validate JSON-LD only. JSON-LD is the format Google recommends and the easiest to maintain, since it lives in one script block instead of being woven through your HTML. If you have both, consider consolidating on JSON-LD.
- Will fixing these warnings improve my rankings?
- Structured data is not a ranking factor in the classic sense, and we won't pretend otherwise. What it does: it makes your entities (brand, authors, products) machine-readable and unambiguous, which is how search engines and LLM knowledge systems connect mentions of your name to the right entity. Errors mean the markup is ignored entirely — that's what's worth fixing first.
- What does an @id-linked graph mean and should I use one?
- Instead of repeating your organization's details in every block, you give each entity an @id (like https://yoursite.com/#organization) and reference it — Article points to its author Person, the Person's worksFor points to the Organization. This linked graph is the current best practice for entity disambiguation because every page reinforces one consistent set of entities.
Related free tools
FAQ Schema Generator
Turn your questions and answers into valid FAQPage JSON-LD structured data. Free, instant, and honest about what FAQ schema does in 2026.
Organization Schema Generator
Generate Organization JSON-LD with logo, sameAs and contact details — the entity markup AI engines use to disambiguate your brand.
Author / E-E-A-T Schema Generator
Create Person JSON-LD for your authors — jobTitle, worksFor, knowsAbout, sameAs — the machine-readable expertise profile behind E-E-A-T.
Now find out what AI models actually say about your brand
These tools make your site readable to AI. The free AI Visibility Checker probes grounded AI models with live web search and shows the real answers, your score, and who wins instead of you — no login required.
Run the free AI Visibility Checker