Compliance registerPROTOTYPE · design preview

The register, writing itself.

Every marking and labeling decision, logged automatically — content-blind. The rows below are real events from this server's evidence log (each mark made in the demo lands here); rows tagged SAMPLE illustrate the event types v1 adds. Tap any row for the full audit record.

Assets logged
—
AI tools seen
—
distinct generator models
Last 7 days
—
events logged
Evidence-set digest
—
sha-256 · binds the pack to these records
Time (UTC)AssetEventAI tool / modelBinding IDStatus
marked — C2PA + watermark applied & loggedlabel applied — visible EU disclosure showncarve-out — label lifted, reason recordedrecovered — provenance restored after strippingchecked — no AI declared, decision recorded
Content-blind by design. The register stores hashes, identifiers, timestamps and decisions — never your images, text, prompts or customer data. That's what makes it safe to keep, and safe to show.