Tool · Voice
Voice Deepfake Detector
Drop a WAV / MP3 / OGG voice clip. We compute pitch jitter, loudness uniformity, ZCR stability and vocoder high-frequency leakage to fingerprint synthetic voices.
Scan settings
Scan mode
Aggressive video scans the full video frame-by-frame; other media uses the selected ensemble mode.
Detection sensitivity
50 · BalancedStrict50 · Balanced (default)High
Frames / images scoring ≥ 50% are flagged as fake. Lower the slider to reduce false positives; raise it to catch borderline cases.