Valeria Metadata Model

Valeria uses a structured metadata model for describing an image, its creator, rights, and declared digital source. Valeria Metadata Compiled (VMC) is the resolved asset-level metadata record. Valeria publishes VMC as YAML for readable inspection and tooling, and serializes the same record as JSON-LD for embedding in C2PA assertions and other JSON-based processing. This page documents the Valeria metadata model and its mappings to standardized fields.

Model structure

The Valeria metadta model currently supports these metadata fields:

PathTypeDescription
metadata.gallery.titleStringHuman-readable gallery title.
metadata.collection.titleStringHuman-readable collection title.
metadata.image.titleStringDescriptive title of the image.
metadata.image.curation_levelEnumerationValeria curation classification: curated, extended, or complete.
metadata.creator.nameStringName of the creator or declared creator.
metadata.creator.organizationStringOrganization associated with the creator.
metadata.creator.identifierURIIdentifier for the creator.
metadata.rights.copyright_noticeStringHuman-readable copyright notice.
metadata.rights.rights_holder.nameStringHuman-readable name of the rights holder.
metadata.rights.rights_holder.organizationStringOrganization associated with the rights holder.
metadata.rights.rights_holder.identifierURIIdentifier for the rights holder.
metadata.rights.license.identifierStringOptional identifier for the licence.
metadata.rights.license.nameStringOptional human-readable licence name.
metadata.rights.license.uriURICanonical URI for the licence.
metadata.rights.rights_statement.identifierStringOptional identifier for a rights statement.
metadata.rights.rights_statement.nameStringOptional human-readable rights-statement name.
metadata.rights.rights_statement.uriURIURI for a rights statement. A rights statement is distinct from a licence.
metadata.source.digital_source_type.labelStringHuman-readable label for the digital source classification.
metadata.source.digital_source_type.uriURIControlled-vocabulary URI for the digital source classification.

The model uses structured terms where identity or semantics matter. Human-readable labels and names support presentation; canonical URIs are used for interoperable identity and controlled-vocabulary values. The VMC 3.0 JSON-LD serialization uses a versioned JSON-LD context that defines the VMC terms, and a JSON Schema that defines the document structure.

The following sections map VMC fields to standardized metadata properties and C2PA assertions. Compact IRIs such as dc:title identify properties in a target metadata or assertion vocabulary, not VMC field names.

Technical noteThe mappings are projections, not claims that source and target fields are semantically identical. Each projection should preserve the meaning of the source field within the vocabulary and structure of the target format.

Conventional metadata

VMC can be mnapped to conventional embedded metadata formats, such as XMP. The following table uses the namespace-qualified property names defined by the relevant XMP schemas.

VMC fieldConventional propertyNotes
asset.identifierdc:identifierStable identifier for the asset.
metadata.image.titledc:titleXMP language alternative; the default language is x-default.
metadata.creator.namedc:creatorXMP sequence containing the creator name.
metadata.creator.organizationphotoshop:CreditOrganization or credit line associated with the creator.
metadata.rights.copyright_noticedc:rightsXMP language alternative; the default language is x-default.
metadata.rights.license.urixmpRights:WebStatementURI of the applicable licence statement.
metadata.rights.rights_holder.identifierxmpRights:OwnerXMP bag containing the rights-holder identifier.
metadata.source.digital_source_type.uriIptc4xmpExt:DigitalSourceTypeIPTC Digital Source Type controlled-vocabulary URI.

The namespace prefixes used in this table are:

dc           http://purl.org/dc/elements/1.1/
xmpRights    http://ns.adobe.com/xap/1.0/rights/
photoshop    http://ns.adobe.com/photoshop/1.0/
Iptc4xmpExt  http://iptc.org/std/Iptc4xmpExt/2008-02-29/

Fields without a listed conventional mapping remain available in VMC but are not assigned an equivalent XMP property by this model.

C2PA provenance

C2PA assertions can carry both descriptive metadata and provenance. The assertion label and the property path are part of the mapping; they should not be conflated with ordinary embedded metadata.

Descriptive assertions

VMC fieldC2PA assertionAssertion property
asset.identifiercawg.metadatadc:identifier
asset.identifierstds.schema-org.CreativeWorkidentifier
metadata.image.titlecawg.metadatadc:title
metadata.image.titlestds.schema-org.CreativeWorkname
metadata.creator.namecawg.metadatadc:creator
metadata.creator.namestds.schema-org.CreativeWorkauthor[0].name
metadata.creator.organizationstds.schema-org.CreativeWorkauthor[0].affiliation.name
metadata.rights.copyright_noticestds.schema-org.CreativeWorkcopyrightNotice
metadata.rights.license.uristds.schema-org.CreativeWorklicense
metadata.rights.rights_holder.identifierstds.schema-org.CreativeWorkcopyrightHolder.identifier
metadata.source.digital_source_type.uricawg.metadataIptc4xmpExt:DigitalSourceType

cawg.metadata uses CAWG/IPTC compact-IRI properties. The stds.schema-org.CreativeWork assertion uses the Schema.org JSON-LD context; for example, license expands to the Schema.org license property. The assertion label stds.schema-org.CreativeWork identifies the assertion type; it is not itself a VMC field.

Rights assertion

VMC fieldC2PA assertionAssertion property
asset.identifierjpt.rightsODRL target
metadata.rights.license.urijpt.rightsdcterms:license
metadata.rights.rights_holder.identifierjpt.rightsODRL permission[0].assigner

The current jpt.rights assertion is an ODRL rights expression. It records the licence, target, assigner, permitted reproduction/display/distribution actions, an attribution duty, and the configured prohibitions. A rights_statement is retained in VMC but is not currently projected into jpt.rights.

Publication provenance

The publication event is represented separately from descriptive metadata and rights:

C2PA assertionPropertyMeaning
c2pa.actions.v2action: c2pa.publishedRecords publication of the asset. The software agent identifies the publishing service.
gallery.valeria.vmcComplete VMC 3.0 documentPreserves the full structured metadata record, including fields without a standardized projection.

metadata.source.digital_source_type is a source classification, not a creation action. A value such as the IPTC URI for “Created using Generative AI” expresses a declared digital source type; it does not by itself prove how, when, or by which system the image pixels were created. The model therefore does not use it to invent a c2pa.created action.