Knowledge Base
Schema.org Implementation for AI Discovery
Digital Karma Federation v8.2 requires the site's real DataCatalog and complete Dataset inventory in the rendered JSON-LD graph on every canonical sitemap page.
Required Entity Types
OrganizationorPerson... the real creator and publisherWebSite... the stable site identityWebPage,Article,FAQPage, or another accurate primary page typeDataCatalog... the canonical catalog from/ai/catalog.jsonDataset... every Dataset declared by the catalog
Recommended Entity Types
Service... offerings and capabilitiesSoftwareApplication... tools and platformsBrand... brand identity- Richer page-specific entities that describe the actual content without competing with the canonical site graph
Implementation
Load the Dataset definitions from /ai/catalog.json. Reuse one stable @id for the site, publisher, catalog, and each Dataset across every page.
Inspect native JSON-LD before adding anything. Preserve richer page-specific fields, merge matching entities onto their canonical identifiers, rewrite references, and add only what is missing.
The primary page entity must connect to the relevant Dataset layer through subjectOf, isPartOf, or an equally explicit relationship. Multiple script blocks are acceptable when they resolve to one coherent graph identity.
Proof Before the Version Claim
Run the rendered Dataset validator against every URL in the root sitemap. It must reject redirects, invalid JSON-LD, missing catalog nodes, missing Dataset nodes, competing identifiers, short Dataset descriptions, and missing page-to-Dataset relationships.
Only a zero-failure crawl can publish /ai/rendered-schema.json with status: passed and coverage: 1. The receipt becomes stale when the sitemap or catalog hash changes.