Metadata
Metadata is descriptive, organizational, or external-identity information associated with media and connected records. It is distinct from source content, file-level technical values, generated assets, and user-specific engagement.
Kinds of metadata
Section titled “Kinds of metadata”Descriptive fields include titles, dates, details, codes, and type-specific values. Relationships connect reusable tags, performers, studios, groups, and galleries. URLs and provider-scoped remote IDs connect local records to outside sources. Custom fields extend the configured schema.
Technical values such as file path, codec, dimensions, size, duration, and hashes belong to files. Playback history, favorites, and similar per-user state are engagement rather than shared descriptive metadata.
Sources and precedence
Section titled “Sources and precedence”Users, scans, providers, scrapers, metadata servers, downloaders, API clients, and extensions can supply values. Provenance can record the source of a field independently, allowing one record to combine values from several origins.
Applying metadata can add, replace, merge, or leave values unchanged according to the workflow and field. A provider match is external identity evidence; it does not make that provider authoritative for every field forever.