← Use cases · Newsrooms & content-trust ops
Audit-trailed verdicts for unsigned media
When a viral clip arrives without provenance, newsrooms need a defensible answer — a verifiable consensus of many detectors, plus a record of how they got it.
The problem
Most media a newsroom evaluates arrives without a signed manifest: reposts, screen recordings, anonymous tips. A single-vendor detection API gives a verdict but no reconstructable evidence, and its definition of "real" can change with a policy update. For editorial trust, "trust our model" is not enough — you want a staked, auditable consensus you can show your work on.
Where dfpn fits
- ▸You evaluate media whose origin is unknown or unsigned
- ▸You need to show your work — a verdict a court, an editor, or a reader can audit
- ▸You do not want the definition of "real" controlled by one counterparty
- ▸You want multiple independent detectors voting, not a single black box
How the flow works
- 01
Submit
Post the media hash and a fee through the SDK. The media stays in your storage; only its hash goes on-chain.
- 02
Consensus
Several independent workers analyze it and reach reputation-weighted consensus on a label and confidence.
- 03
Cite
Attach the on-chain audit trail — assigned workers, commits, reveals, contributing models — to your editorial record.
Honest caveat
dfpn is a detection signal, not an oracle. Adversarial media can bypass any detector; treat the verdict as strong evidence, not proof, and pair it with provenance when the producer cooperated.
Other use cases
Put it to work
Wire verdicts into your pipeline, or read how a verdict is produced end to end.