Audio Provenance
Prove a recording carries its creator’s mark.
ObraVera embeds tamper-evident credentials into audio recordings, carried by a watermark designed to survive common compression and format conversion. Prove provenance at any point in the distribution chain.
AI voice cloning is here. Your members need proof their recordings are real.
Voice Cloning
Voices cloned without consent for commercial use.
Metadata Stripped
No provenance once metadata is stripped.
Creators need credentials that live inside the audio itself.
Acoustic verification — without transferring any file.
Hold your phone near a speaker playing back an ObraVera-authenticated recording. The verifier listens for the embedded mark — no file transfer needed. No audio is stored.
- Designed for speaker and broadcast playback
- Designed to survive common platform processing
- Typically ~20 seconds of capture — no file access needed
The Hidden Signal
Inaudible to humans. Unmistakable to the detector.
The audio waveform below looks normal. The ObraVera watermark is embedded inside the signal at a level imperceptible to the ear. Watch the detector sweep and reveal the hidden credential layer.
If one layer is stripped, others remain.
Acoustic Watermark
Inaudible, embedded in the audio signal itself
Acoustic Fingerprint
Detects tampering and unauthorized edits
Cryptographic Signature
Mathematical proof of authenticity
Registry Recovery
Credentials recovered even when metadata stripped
Designed to survive everyday distribution, within stated limits.
At its default strength, audiowmark, the open-source watermark ObraVera builds on, is documented as recoverable after MP3 or Ogg encoding at 128 kbps or higher; ObraVera uses a stronger default. Lower bitrates and repeated re-encoding weaken the mark, and it is not designed to survive re-synthesis by a voice-cloning model. In internal testing, it survives -14 LUFS loudness normalization. Our own measurements across codecs, bitrates and strengths will be published here.
Supported Formats
Audit-Ready
Proof you can show a client — or a union.
As AI transparency obligations converge, studios and unions need evidence and workflows — not just trust.
Signed Custody Chain
Every credential is cryptographically signed at the point of creation — immutable, timestamped, and registry-anchored.
Local Verification
Verify any file without uploading audio. Your content stays yours — verification runs without transferring the recording.
Receipts When You Need Them
Even when metadata is stripped, registry-anchored credentials provide documentary evidence for disputes and licensing.
Who It's For
Built for the people who create
Voice Actors
Prove which recordings are authentically yours, even when your voice is cloned.
Musicians & Studios
Embed tamper-evident credentials in every track. Verify provenance across distribution.
Performers' Unions
Give your members the tools to prove their work is real. Protect their livelihoods.