Provenance research traces the history of an object’s ownership from its creation to today, through sale catalogues, dealer records, letters and archives. AI speeds up the searching and reading; a provenance researcher verifies every link and draws the conclusions.
All Museums, Heritage & Auction →
A provenance file is built one source at a time: a label on the back of a canvas, a lot number in a 1937 sale catalogue, an entry in a dealer’s stock book, a name in a restitution claim. The sources are scattered across libraries and databases, written in several languages, often digitised as images rather than text. The slow part is not judgement; it is finding and reading.
So the system does the finding and reading. Given an object’s record and what is already known, it searches digitised sale catalogues, archives you have access to and the databases a researcher would check, such as the German Lost Art Foundation’s database, the Art Loss Register and the Interpol stolen works of art database where you hold access, and returns candidate matches. Each comes with the source, the page and the reason it might match, for the researcher to confirm or reject.
The output is a timeline in which every step is either cited and confirmed by a person, or marked as a gap. Gaps are never smoothed over, least of all in the 1933–1945 period, where the Washington Conference Principles set expectations for research and transparency.
We test on provenance files your researchers have already completed. The system starts from the same object record and we compare what it finds with the finished file: which sources it located, which it missed, and whether any candidate it proposed was wrong. That tells the team where to trust it as a search tool and where to keep searching by hand.
For auction houses, the same research supports due diligence on consignments, alongside the anti-money-laundering checks on parties to a sale that art market participants must carry out under the EU AML directives and the UK Money Laundering Regulations. We build the research and record-keeping; the compliance decision belongs to your compliance officer.
what is already known
catalogues · archives · databases
source, page and reason
or rejects each lead
cited steps and marked gaps
1933–1945 or a loss-database hit
The study of an object’s ownership history, from when it was made to the present day. Researchers trace each owner and transfer through sale catalogues, dealer records, inventories, letters, photographs and labels on the object itself. It establishes authenticity and legal title, and identifies objects that may have been looted, stolen or illegally exported, especially during the Nazi era and in colonial contexts.
It can do much of the searching and reading: matching names and objects across digitised catalogues, transcribing and translating documents, and checking loss databases. It cannot weigh conflicting evidence or decide what a gap means. We build it to cite every lead and mark every gap, and a provenance researcher confirms each step and draws the conclusions.
Commonly used ones include the German Lost Art Foundation’s Lost Art database, the Art Loss Register, the Interpol database of stolen works of art, digitised sale catalogue collections held by research libraries, and national and institutional archives. Access terms differ, and some are subscription-based. We integrate only with sources you are entitled to use.
What we built, what stage it is at, and a link to the live thing wherever there is one.
yourOS is an operating system with an AI at its heart. You talk to it like a person, out loud or typed. It makes your calls, plans your days and works on your projects with a crew of agents, on your own private server. Our own product, built from the first commit; the waitlist is open.
Website, backend and app assets for Sonora, a free AI sound-therapy app on iOS and Android published by Cognitive AI.
Brand, website, prototype, apps on five platforms, the AI backend and the algorithm-intelligence pipeline for Podswap, a social growth engine that runs without a human in the loop.
In closed beta: Roxy, a private AI assistant with permanent memory running on infrastructure the user owns. Brand, site, product, backend and conversational design.
Live at coherence.com: brand identity generated with AI agents we directed, the motion system and website for the group behind Sonora, AboutAI, WakeUp and MudShadows. Agent work remains in build.
Live at domainbooth.com for DomainBooth, a premium domain brokerage: a new animated mark, every page redesigned and rebuilt with designed mobile layouts, and a custom CMS for its sales record, featured names, press and enquiries.
Bring a completed provenance file; we will run the system from the same starting record and show what it finds and misses.
30-minute technical call · no deck