HSI is built to be inspected. We publish our methods, our model versions, our evidence integrity, and our mistakes.
Our full pipeline — from capture to publication — is documented and public.
Every classifier version is recorded so findings can be reproduced and audited.
Archived evidence is hash-sealed; integrity can be independently verified.
When we get something wrong, we publish the correction and the reason.
We report a live traceability score across all published findings — and we hold ourselves to it.