Structured data workflow
JSON-LD: truthful structured data and safe script integration
Structured data describes the content a page actually provides. More fields do not make invented or inconsistent claims acceptable.
Content updated: · Maintainer and corrections
Choose the entity your page represents
Article describes an article, BreadcrumbList an ordered navigation trail, and WebSite the site-level entity. Use actual names, dates, authors and URLs. Omit unavailable optional details rather than inventing a review, rating, publisher or publication date.
Schema.org vocabulary and a search engine's supported features are different contracts. Google removed the sitelinks search box in November 2024; adding an old SearchAction template does not restore it. Check current feature documentation instead of assuming a generator installs a search feature.
Valid JSON still needs safe HTML embedding
A JSON string can contain </script>; placing raw serialized JSON into HTML can prematurely close a script element. Use the generated escaped script output for that context, not a hand-built string or a generic HTML entity encoder. Parse exported JSON separately to check that escaping preserves the intended data.
The local preview displays text and does not execute your markup. After integration, inspect the rendered page, test current required properties with the appropriate validator and monitor the deployed result. Passing JSON parsing is not full semantic validation, and even correct structured data does not guarantee a Google rich result.
Data provenance