Schema Validator for SEO and GEO

The Schema Markup Checker works as a schema validator, schema tester, structured data checker, and JSON-LD checker in one — and runs type-specific checks for formats like FAQ schema, Article, Product, and LocalBusiness. It supports two input modes so you can check your schema whether your page is live or still in development.

Validate by URL fetches your live page and extracts all <script type="application/ld+json"> blocks automatically. Enter the full page URL (including https://), click Validate, and the tool parses every schema block it finds, returning a separate validation report for each one. This is the fastest way to audit an existing page and confirm that your CMS or developer has implemented schema correctly.

Paste JSON-LD lets you validate schema you write yourself before publishing it. Paste your JSON-LD object (including the @context and @type declarations) into the text area and click Validate. The validator accepts both single schema objects and JSON arrays containing multiple blocks. This mode is particularly useful when testing changes to existing schema or validating schema generated by our Schema Generator before deploying it.

After validation, you receive a schema score (0–100) alongside a per-block breakdown. Each block shows its type, a valid/invalid badge, a toggle to view the raw JSON, and itemised errors (red), warnings (amber), and suggestions. Fix errors first — they represent missing required properties or invalid values that prevent search engines from processing your schema entirely.

Schema markup is structured data embedded in a webpage that describes its content in a machine-readable format. The dominant implementation today is JSON-LD (JavaScript Object Notation for Linked Data): a <script type="application/ld+json"> block placed in the page head or body containing a JSON object that follows the vocabulary defined at schema.org. The schema.org vocabulary is maintained jointly by Google, Microsoft, Yahoo, and Yandex and covers hundreds of types — from Article and Product to Event, Recipe, and LocalBusiness.

Each schema type has a defined set of properties, some required and some recommended. An Article schema, for example, should include headline, author, datePublished, and publisher. A Product schema needs a name and can include offers, aggregateRating, and brand. A FAQPage requires a mainEntity array of Question objects, each with an acceptedAnswer. When these properties are present and correctly formatted, search engines and AI platforms can parse the data with high confidence.

For traditional SEO, valid schema markup unlocks rich results in Google Search: FAQ dropdowns, star ratings, How-To cards, event details, and recipe metadata that increase click-through rates. But the impact for AI visibility is arguably more significant. When ChatGPT, Perplexity, or Gemini processes a page for inclusion in an AI-generated answer, structured data provides explicit signals that bypass the uncertainty of natural language inference. A page with valid Article schema telling the engine that the author is Dr. Jane Smith, published on March 15 2026, about keyword X — is far more likely to be cited accurately than an identical page with no schema at all.

LocalBusiness schema is particularly valuable for brands with physical locations. Including address, telephone, openingHours, and geo coordinates lets AI search engines surface your business in location-based queries without relying solely on third-party directory data. Similarly, Product schema with complete offers data (price, currency, availability) helps AI engines answer purchase-intent queries with accurate information directly from your site, reducing dependency on aggregators.

Getting schema right is not a cosmetic detail — it directly affects both how your pages perform in Google (SEO) and how they are represented by AI answer engines (GEO). Checking your markup before and after you publish is the only way to know the structured data you shipped is the structured data search engines and AI models actually read.

For SEO, valid structured data is what makes your pages eligible for rich results — FAQ dropdowns, star ratings, How-To steps, breadcrumbs, and product details that take up more space in the search results and lift click-through rates. A single missing required property (an Article with no author, a Product with no name, an invalid ISO date) can silently disqualify a page from those rich results even when the rest of the markup looks fine. Checking your schema catches those disqualifiers before they cost you visibility.

For GEO, AI engines like ChatGPT, Perplexity, Gemini, and Google AI Overviews lean on structured data to extract facts — author, date, entity type, price, rating — without inferring them from prose. Clean, valid schema reduces ambiguity, which means more accurate citations, correct attribution, and a higher chance your page is surfaced in AI-generated answers. Invalid or conflicting schema does the opposite: it forces the model back to guessing, and guesses are where misattribution and lost citations come from.

This checker validates your markup against the official schema.org vocabulary and Google's structured data guidelines — the same official standards search engines and AI crawlers use. It then returns our schema score: a 0–100 composite built by Geology that weights validity, completeness, and whether the right schema types are used for your content, giving you a single number to track alongside the specific fixes.

AI search engines like ChatGPT, Perplexity, and Gemini generate answers by extracting and synthesising information from crawled web pages. When a page includes valid structured data, these engines don't need to infer facts from prose — the facts are declared explicitly in a machine-readable format they are designed to parse. This distinction matters more than many SEO professionals realise.

Consider a product review page. Without schema, an AI engine must parse the prose to determine the product name, rating, reviewer, and publication date — tasks it handles reasonably well but not perfectly. With valid Review schema including itemReviewed, reviewRating, and author, those facts are declared with zero ambiguity. The AI engine processes them as structured inputs rather than probabilistic extractions from natural language. The result is more accurate citations, less hallucinated metadata, and higher confidence that the page will be surfaced for relevant queries.

Google's approach to AI Overviews relies heavily on structured data. Pages with well-implemented FAQPage schema are extracted as Q&A pairs that can appear directly in AI Overview responses. HowTo schema provides step-by-step instructions that AI engines can present in structured format. Breadcrumb schema helps AI engines understand site hierarchy and content relationships — making it easier for them to determine whether your page is an authoritative source on a topic or a peripheral page on a tangentially related site.

From a technical SEO and GEO perspective, schema markup is the most direct lever you have over how AI engines interpret your content. Fixing invalid schema — particularly missing required properties on Article, Product, and FAQPage schemas — is consistently one of the highest-ROI improvements we see in AI visibility audits. The cost is low (a few JSON-LD properties), the impact is measurable, and the changes take effect as soon as the page is re-crawled.

These are the five schema errors we encounter most frequently when auditing sites for AI search visibility:

• Missing required properties. The most common issue. Article schema without author, Product schema without name, FAQPage without mainEntity. When required properties are absent, most validators treat the schema as invalid and AI engines fall back to inferring the data from unstructured content.
• Invalid date formats. Date properties like datePublished and dateModified must follow ISO 8601 format (e.g. 2026-03-15 or 2026-03-15T10:00:00+00:00). Dates formatted as "March 15, 2026" or "15/03/2026" are invalid and will cause parsers to discard the property entirely.
• Duplicate or conflicting schema blocks. Publishing two Article schema blocks on the same page — common when a CMS adds schema automatically and a developer adds another manually — creates conflicting signals. Validators flag this, and AI engines may use the wrong block or average across both, producing incorrect metadata.
• Deprecated schema types. Types like NewsArticle without proper isAccessibleForFree declarations, or using itemtype microdata instead of JSON-LD, can produce unexpected results as engines deprioritise older implementations.
• Schema that does not match page content. Adding Product schema to a blog post, or FAQPage schema to a page with no actual questions and answers, violates Google's structured data quality guidelines and can lead to manual actions. AI engines are increasingly able to detect schema-content mismatches, and pages flagged for spammy markup see reduced citation rates.