An agentic discovery-intelligence platform that turns the world's known unknowns into active Hunts — mapping what exists, what's missing, and what might exist, then orchestrating AI agents and human investigators to follow the evidence, pursue the leads, and find personal or historic treasures potentially lost to time.
Before anything is hunted, the domain is mapped three ways — what is documented and located, what is documented and missing, and what the record implies should exist but was never catalogued. The gap is the asset.
Each pursued gap opens as a Hunt with a defined question, a live evidence set, and a standard of proof. Agents and human investigators work the same case file, and the Hunt closes only on evidence — including a documented dead end.
Sources, inferences, and contradictions are held as a graph rather than a summary. Any conclusion can be walked back to the documents it rests on, and confidence moves when the evidence moves — not when the model feels certain.
Agents run the search at a scale no research team could, then hand off. Consequential calls — what to pursue, what to trust, what to act on in the physical world — stay with the investigator, on the record.
Interactive mockups from the platform in development — the Hunt console where investigations are run, and the evidence board where a single Hunt is worked and adjudicated.
In active development. If you work in archives, provenance, recovery, or investigative research — become a design partner and help shape it.