structured-content — semantic metadata that is inherited.

Architecture

Classes/
├── Domain/          # Annotation models, relationship types, contracts
├── Application/     # Context resolution, relationship services
├── Infrastructure/  # Repositories, JSON-LD builders, context resolvers
└── Presentation/    # Controllers, event listeners

Database tables

Schema.org generator stack

All generators in Classes/Infrastructure/SchemaOrg/ implement SchemaGeneratorInterface, are registered via the structured_content.schema_generator service tag and rendered in priority order (higher = earlier in the JSON-LD block).

Builders and traits

Application/Service/PersonSchemaBuilder.buildFromRow($personRow) emits Person with worksFor[], sameAs[] (with a Label|URL pipe format), knowsAbout, knowsLanguage and image. PersonRoutingArgumentTrait resolves the person slug from the _routingArguments of the Person plugin detail view. PodcastFeedCacheReaderTrait reads cached feed items for podcast schemas without a re-fetch. SanitizesSchemaTrait strips empty and non-JSON-conformant values. ContextAndRelationshipsTrait attaches AI context and relationships to the schema output.

Repositories

AuthoredPagesRepository.findByPersonUid finds pages that link to a person via tx_structuredcontent_author_records (for Person.hasPart). CategoryRepository.findByPageUid reads sys_category records via sys_category_record_mm (for Article.articleSection). EpisodeGuestsRepository serves bulk episode_slug → Person[] for PodcastEpisode.actor.

Article schema extensions (AI Disclosure, Reading Time, Language Fallback)

AI Content Disclosure (v1.5.10+)

For AI-tagged hero images, the Article schema contains image: {creator: {SoftwareApplication, name}, creditText: "AI-generated"}. The EU AI Act Article 50 disclosure is therefore anchored directly in the JSON-LD. Edited through the TCA fields from moselwal/content-provenance (tx_provenance_origin, tx_provenance_ai_system).

Reading Time (v1.5.9+)

The Article schema carries wordCount and timeRequired (ISO-8601 PT*M) — byte-identical to the frontend reading-time badge (200 words per minute). Unifies the reading-time display between markup and user view.

Language fallback in schema generators (v1.5.7+)

Translation-aware: default-language records are loaded and EN translations are overlaid. The schema output keeps working under fallbackType: fallback without gaps — incomplete localisations no longer cause missing schema markup.

Relationship type mappings (Schema.org)

• replaces → isBasedOn
• belongs_to → isPartOf
• related_to → relatedLink
• explains → about

sys_category mapping

Extends sys_category with a slug field (auto-generated from the title) for filter URLs and an RTE description for category hero blocks. The Article JSON-LD emits the category as articleSection.

CLI commands

vendor/bin/typo3 structured-content:missing-context   # Find pages missing AI context
vendor/bin/typo3 structured-content:export-relations   # Export relationships as JSON
vendor/bin/typo3 structured-content:validate-channels  # Validate channel configurations

Configuration

AI context fields are added to the TYPO3 page and content element forms. Configure default context values and inheritance behaviour through site settings.

Requirements

• PHP 8.3+
• TYPO3 14.0

Optional dependencies

The codebase follows a strict DDD 4-layer separation; layer violations are blocked in CI via deptrac.

Where we use this …

This package is part of the platform line for AI-Ready CMS and AI-Ready Commerce — as the semantic layer that makes content readable for retrieval systems and agents. Managed variant: AI-Ready CMS as a Service.