← STACK/HUB · Provenance / Check · RUBRIC v1.1

Can you prove what's AI and what isn't?

Inspect a file for AI-provenance metadata, then score your organization's readiness against the EU AI Act's transparency and labeling requirements.

Why now: Anthropic began issuing machine-readable watermarks on Claude output to meet new EU AI transparency rules, with older models expected to follow. Every brand, publisher, and agency shipping AI-touched content into the EU now has a disclosure obligation — most have no way to check their own output.

01Metadata Inspector

Drop an image. This reads its embedded EXIF/XMP metadata directly in your browser — nothing uploads anywhere — and flags any AI-provenance or content-credential fields it finds (e.g. C2PA claims, "Digital Source Type", generator tags).

Drop image here

or click to browse — JPG, PNG, WEBP

02Disclosure Readiness

7 fields, scored. Are your content pipelines actually set up to meet transparency obligations, or just aware of them?

7 fields · takes ~2 minutes

Paid pack — not available

The local marker scan and disclosure builder are free. Bulk scanning and managed monthly re-checks are not implemented. The historical $9/mo proposal is not for sale; scope and pricing require approval.

Paid offer unavailable