Glossary
Metadata
Data *about* a document — its category, dates, parties, value, classification, owner — distinct from the document content itself.
Metadata
Metadata is data about a document — its category, dates, parties, value, classification, owner, tags, retention policy, version — distinct from the document content itself.
In Papyrus, metadata comes from three sources:
- System-generated: upload date, uploader, file hash, file size, version number
- AI-extracted: invoice fields, contract parties, dates, KRA PIN
- User-set: tags, classification labels, custom field values, notes
Metadata is what makes search work — both keyword search (matches in metadata are weighted higher than matches in body) and filters (date range, file type, classification). Metadata is the spine of the analytics layer.
Good metadata practice: don't think of it as “extra work after upload”. The right metadata is captured at upload (often by AI), not retrofitted later.