Content Credentials
Checks for supported C2PA provenance records, including declared AI generation, AI assistance and digital capture claims.
Inspect camera metadata, audio, timelines, AI disclosures, provenance and Content Credentials, and receive a Human-Origin Confidence Score (HOCS).
Questions are randomized and designed so you do not need to disclose personal information.
Vloggi Confirm combines capture, participation, consent, custody and media-provenance evidence into a versioned Human-Origin Confidence Score. An arbitrary imported file has less evidence available than video collected through Vloggi's controlled capture workflow.
Checks for supported C2PA provenance records, including declared AI generation, AI assistance and digital capture claims.
Examines camera and device markers, audio and metadata tracks, timestamps, location presence, codecs, frame rates and editor/export markers, then binds the assessment to the file's SHA-256 fingerprint.
Where enabled with consent, checks for recognizable speech, timing continuity and whether a managed capture contains the randomized spoken challenge. Speech is supporting evidence, not voice identity or proof that a voice is natural.
Separates evidence found from cautions. Missing evidence does not mean fabricated media, and a strong result does not certify that the events shown are true.
For an imported file, Confirm assesses the provenance and media evidence available in that file. In Vloggi's managed capture workflow, it can also evaluate human participation, consent, device interaction and chain of custody. It does not claim that every event depicted is true.
It can identify supported Content Credentials and declared AI-generation or AI-assisted provenance. A missing disclosure does not prove that a video was made by a person.
That is treated as missing provenance, not proof of manipulation. Many genuine cameras, apps and older files do not yet attach C2PA credentials.
A Human-Origin Confidence Score of 100 is reserved for eligible media recorded through a managed Vloggi enterprise capture workflow. It requires a verified, signed-in participant using a registered and attested Vloggi app on a supported device, multiple responses to randomized prompts, bound consent and a validated chain from capture to upload, with no contradictory evidence. Vloggi can configure this maximum-assurance workflow for organizations, but a HOCS of 100 is issued only when every required signal validates. It does not mean a 100% probability or guarantee. Talk to Vloggi about enterprise verification →
HOCS is a 0–100 evidence score describing the strength of available evidence that media originated through human participation in a documented process. The bands are Weak (0–39), Limited (40–64), Moderate (65–79), Strong (80–89), Very strong (90–96) and Exceptional (97–100). HOCS is not a percentage probability, identity proof or guarantee that the events shown are true. The underlying model weights remain proprietary.
Only after you explicitly consent. Confirm uploads an encrypted temporary copy for automated metadata and media-structure analysis, then deletes it immediately when processing finishes. A one-day storage lifecycle is the failure-safe if processing is interrupted. Vloggi retains the assessment evidence and SHA-256 fingerprint, not the source video.
It is a repeatable identifier calculated from every byte of the file. If the file changes, its fingerprint changes. It does not identify the person who created the video.