The Provenance Engine.

AI-powered extraction, reconstruction, and risk detection — all in one place.

How It Works

From document to provenance report in six simple steps.

Step 1

Upload your documents

Certificates, invoices, catalog pages, auction records, museum labels — any provenance documentation.

Step 2

AI extracts the data

Names, dates, locations, ownership changes, and exhibitions — all parsed automatically. People and institutions are matched against Getty ULAN, Wikidata, and other authority databases.

Step 3

Link to an artwork

Connect documents to an existing artwork in your collection, or create a new one from the extracted details.

Step 4

Review and accept events

AI suggests provenance events from each document. Duplicates are automatically flagged. Accept, edit, or reject suggestions to build your chronological timeline.

Step 5

See risks and gaps

Maros automatically detects ownership gaps, conflicting dates, and other provenance red flags on your timeline.

Step 6

Generate a report

Export a research-grade provenance summary with full source citations and risk assessment.

Ready to see it in action?

Upload your first document and experience the power of automated provenance.