Blog
Notes on honest detection
Short, practical notes about detector scope, uncertain verdicts, evidence reports, and review workflows.
Launch notes
How we scope a trustworthy AI image detector
Why we prefer honest supported-image claims over vague all-image promises.
Trust
Why uncertain results matter in image detection
A detector should avoid overconfident claims when metadata or visual evidence is mixed.
Workflow
How teams can review suspicious image uploads
Use scan IDs, evidence reports, and review status to keep moderation decisions traceable.
Longer write-ups land here as the public beta progresses — model updates, benchmark methodology, and detection deep-dives.