Schema Markup Validator

Three kinds of markup

Markup is written in three ways: JSON-LD as a separate block, microdata through the itemscope and itemprop attributes, and RDFa through typeof and property. Live sites use all three, and older projects lean toward the second and third.

The validator looks for all three and shows what it found in each kind. This matters: if you check only JSON-LD, a site with microdata looks like a site with no markup at all.

Fields by type

Each type has required fields, without which a rich snippet cannot be built, and recommended fields that make it richer. The tool checks both and explains for each type what the markup is for: product price and availability, review stars, breadcrumbs in place of a URL.

For Product and Article, Yandex and Google differ in their requirements. Yandex wants an image for a product, and a publication date and author for an article, while Google treats these as optional. The tool shows such differences on a separate line, so markup built for one search engine does not turn out incomplete for the other.

Broken JSON

An extra comma or an unclosed quote inside a JSON-LD block means the search engine drops the whole block, silently. Nothing changes on the page, but the markup no longer exists for search.

The validator reads every block, and when one fails to parse it names the reason in plain words and shows the block itself. Hunting for a missing comma in someone else's code takes a long time otherwise.

Worth keeping in mind

• markup describes what is visible on the page: marking up hidden text or third-party ratings is not allowed;
• having markup does not promise a rich snippet, as the search engine decides on its own;
• product markup belongs on the product page and article markup in the article; home pages usually have neither.

The page is read once, with the same pace of requests to third-party sites as the other checks.