Structured Data Contract¶

Homebase treats structured metadata as a shared machine-readable description of the visible site, not as markup for one search engine.

Preservation rule¶

Valid structured information is preserved by default. A metadata field or Schema.org entity is removed only when it is inaccurate, misleading, invalid, or unrelated to visible page content.

The absence of a rich result in one search engine is not a reason to remove otherwise accurate semantic information.

Consumer dimensions¶

Every structured-data capability is evaluated independently across three dimensions:

Dimension	Question
Semantic validity	Does the type and its data accurately describe visible content using the shared vocabulary?
Consumer support	Does a specific search engine or application document a feature based on the data?
Experimental interoperability	Could other crawlers, agents, or tools consume the data even without a documented presentation feature?

Consumer-specific support can change without changing the underlying meaning of the structured data.

Accuracy requirements¶

Homebase metadata must:

match the canonical URL and primary purpose of the page;
describe content visible to visitors;
use real publication and modification dates;
omit unknown facts instead of manufacturing fallback values;
distinguish generatedAt from publishedAt and modifiedAt;
preserve stable entity identifiers across pages;
connect related entities explicitly;
serialize configurable text and URLs as valid JSON;
distinguish site generation time from content modification time.

Validation¶

Homebase automates structural validation:

Every generated JSON-LD block must parse as JSON.
Fixture builds cover profile, integration, video, review, and shop output.

Release verification uses external tools separately:

Schema.org Validator reviews general vocabulary and property relationships.
Consumer-specific tools check only the features that consumer documents.

Passing a consumer-specific validator is not a guarantee of presentation, and failing to qualify for a presentation feature does not invalidate otherwise accurate Schema.org metadata.

Page graph model¶

Homebase emits one connected JSON-LD graph per generated HTML page:

WebSite and Person use stable site-level identifiers.
The home route uses ProfilePage as its primary page entity.
The shop route uses a page-scoped CollectionPage.
Page-specific item lists and breadcrumbs use identifiers based on their canonical route.
Optional FAQ, video, review, and product entities reference the stable page and creator entities.

Complete objects are serialized in JavaScript rather than assembled as JSON fragments in templates.

Support matrix¶

Status last reviewed: 2026-07-19.

Entity or format	Visible source	Semantic purpose	Consumer-specific status	Experimental interoperability
`WebSite`	Site title, description, language	Identifies the site and publisher	Google supports `WebSite` for site-name understanding; the Sitelinks Search Box was retired in 2024	General site identity for Schema.org consumers
`ProfilePage`	Profile card and biography	Identifies the canonical creator profile page	Google documents `ProfilePage` primarily for creator profiles in community and discussion contexts; presentation is not guaranteed	Stable profile-page identity for graph consumers
`Person`	Name, biography, avatar, roles, socials	Describes the creator and connects identity signals	No single Google rich result is guaranteed	Core creator entity for search, agents, and linked-data tools
`CollectionPage`	Shop heading and collections	Identifies the shop route as a collection page	No dedicated Google rich result	Gives the shop a canonical page identity instead of reusing the profile page
`ItemList`	Visible links or shop cards	Describes ordered page items	Used as supporting structured data; presentation depends on context	Preserves ordering and item relationships
`Article`	Recent-content records	Describes externally published creator content	Article features generally require the article's own visible page and policies	Connects the profile to creator-authored content
`FAQPage`	Visible FAQ questions and answers	Expresses structured Q&A	Google removed FAQ rich results in May 2026	Remains valid Q&A for other Schema.org-compatible consumers
`BreadcrumbList`	Visible route hierarchy	Describes navigation hierarchy	Google supports breadcrumbs when the hierarchy is meaningful	General route context
`SpeakableSpecification`	Profile bio and FAQ answers	Marks concise text-to-speech candidates	Google's implementation is a US English topical-news beta	Preserved for other or future speech consumers
`VideoObject`	Visible YouTube embeds	Describes video title, URLs, image, creator, and known upload date	Google video features focus on watch pages and require feature-specific fields	General video semantics even when a homepage is not an eligible watch page
`Review`	Visible testimonials	Describes attributed testimonial text and optional ratings with `itemReviewed` referencing the creator	Google review snippets do not support `Person`	Machine-readable creator reviews when truthful and visible
`AggregateRating`	Visible numeric testimonial ratings	Summarizes actual rated testimonials and identifies the creator through `itemReviewed`	No Google `Person` review-star eligibility	General aggregate semantics; omitted when no numeric ratings exist
`Course`	Visible course cards	Identifies educational offerings	Google phased out Course Info search presentation in 2025	Course meaning remains valid beyond that presentation feature
`Book`	Visible book cards	Identifies books and their creator	Google announced the phase-out of Book Actions support in 2025	Book semantics remain valid for other consumers
`Service`	Visible service or consultation cards	Identifies creator services and offers	No dedicated Google service rich result	Useful service semantics for other consumers and agents
`Product` + `Offer`	Visible product cards, URLs, and prices	Describes products and known offers	Google product features require visible, policy-compliant product information and may require additional fields	General commerce semantics for compatible consumers
`EducationalOccupationalCredential`	Visible or configured credentials	Describes creator credentials	No dedicated Google rich result	Identity and qualification context
`InteractionCounter`	Visible or fetched audience statistics	Describes follower and publishing counts	No guaranteed search presentation	Quantitative social-proof context
`llms.txt` / `llms-full.txt`	Profile, links, FAQ, products, and content summaries	Provides optional plain-text machine context	Google explicitly states that it ignores LLMS.txt for Search visibility	Emerging convention used by some documentation and developer tools

Interpretation rules¶

Semantic validity does not imply a guaranteed search presentation.
The absence or retirement of a presentation feature does not invalidate truthful Schema.org data.
Search-engine requirements are checked independently from general Schema.org validity.
Experimental interoperability is documented as a possibility, never as a ranking or citation guarantee.