Security

Every document is processed securely and is never used to train external models. Access to client files is role-based, and every review, flag, and decision is logged in a full audit trail — traceable end to end.

A detailed security overview, including infrastructure and compliance posture, will be published here before general availability.