Reading the description, creator disclosures, and viewer signals…
This takes a few seconds. We never guess about pixels.
Signals we read
Observed directly from this video's public text and audience. This is a signal-based likelihood, not forensic proof.
How to read this
This is a signal score, not a detection verdict. It measures how strongly the public record, the creator's disclosures, the description, and what viewers report, points toward AI or manipulation. A high score means those public signals are strong. A low score means they are absent, which is not the same as proof a video is authentic.
Two layers. This page is the signal layer: what the public record says. Determining whether the footage itself is synthetic is the forensic layer: a frame-level, voice, and editing review done by an analyst. That is the deeper analysis below.
Deeper forensic analysis
Go beyond the public signals.
A human-reviewed analysis examines the parts no metadata can reveal: the voice, the frames, the edits.
Voice Consistency
Cloning, splicing, and synthetic-speech detection.
Visual Inconsistencies
Frame artifacts, lighting and anatomy anomalies.
Editing Anomalies
Re-timing, compositing, and manipulation patterns.
Manipulation Synthesis
An analyst's combined judgment across every signal.
Request deeper analysis