ORAVYS measures the acoustics of a recording: signs of synthetic speech, editing traces, and whether one speaker is talking. It reports on the audio, not on the person. One recording. One report.
Record a conversation directly in your browser or upload an audio file. No app download required.
The engines measure acoustic properties of the audio: spectral consistency, codec and encoding traces, pitch and noise statistics.
Receive a report with synthetic-speech indicators, editing traces and acoustic provenance, each with its uncertainty.
A voice message or a call recording seems odd. The check looks at the audio itself: signs of AI generation, editing, or several speakers. It does not judge what the speaker says, thinks or feels.
Before signing a deal, check that the recording is what it claims to be: a real human voice, unedited, one speaker.
Verify whether a voice message or call is from a real person or AI-generated. Deepfake voices can be hard to tell apart by ear. The report gives a score with its uncertainty, not a certainty.
Check a recording before you share it: level, background noise and pitch statistics, measured on the audio.
Acoustic indicators of whether the recording is a human voice, unedited and from one speaker, each with its uncertainty.
Locate where the recording changes: splices, codec switches, re-encodings and segments that do not match the rest of the audio.
Level, background noise and pitch statistics, so you know how far the measurement can be trusted.
Forensic analysis of audio encoding, spectral consistency, and synthesis artifacts. The result is a score with its uncertainty, not a yes-or-no answer.
Voice Intuition processes biometric voice data under GDPR Article 9(2)(a) with your explicit consent. Before any analysis begins, you confirm that you understand and agree to the processing.
Also: Single scan $9.90, Advanced $99/mo, Enterprise from $2,900/mo by contract. View all plans
Audio is processed in EU data centers and deleted after analysis; any temporary copy is purged automatically within 3 days. Reports are kept 30 days.
ORAVYS combines several independent acoustic measurements. Results are probabilistic indicators with their uncertainty, not definitive judgments, and they describe the audio, not the person.
Yes. Voice analysis involves biometric data processing under GDPR Article 9. We rely on your explicit consent (Art. 9(2)(a)) before any analysis begins. You can withdraw consent and delete your data at any time.
The engines score a recording for signs of AI-generated voices, cloned speech, and manipulated audio, using codec artifacts, spectral consistency, and synthesis patterns. The result is a score with its uncertainty, not a guarantee.
Multi-signal acoustic measurements. One report on the audio, with its uncertainty.
Start Your Free Analysis