Evidence state, not file storage
The target is to preserve why a claim was believed, what evidence supported it, what contradicted it and when it was last revalidated.
The long-term Epistamate direction is institutional research that carries forward claim identity, provenance, review history, applicability, recency and contradictions. That is a different problem from simply giving a model more context.
The research product is early access. The long-lived cross-topic Knowledge layer remains under canonical rebuild rather than being claimed here as finished.
The target is to preserve why a claim was believed, what evidence supported it, what contradicted it and when it was last revalidated.
Model suggestions, researcher decisions and evidence authority remain distinct so governance does not reduce to “a human was in the loop”.
The current product is a Windows desktop application with local canonical persistence and explicit provider/network permissions, not a finished team/cloud platform.
Same-topic continuation and reassessment exist; durable cross-topic canonical Knowledge is being rebuilt and is not claimed as complete today.
We are looking for workflows where source support, provenance, researcher judgement and repeatability matter more than faster prose.
Start a conversation →