Create a unique voiceprint for identity verification
/voiceid/enroll endpoint creates a unique, non-reversible voiceprint (embedding) for a user or speaker. This voiceprint can later be used to verify if a voice sample belongs to the same speaker.
"user_12345" or "john.doe@example.com""titanet-large"| Code | Description | Solution |
|---|---|---|
| 400 | Validation error | Check audio duration (>= 10s) and format |
| 401 | Unauthorized | Verify API key is valid and active |
| 422 | Unsupported format | Convert to supported audio format |
| 500 | Processing failed | Retry or contact support if persists |
Recording Quality
Storage & Security
Multiple Enrollments
Vector Storage
Upload a clean voice sample (>= 10 seconds) to generate a voice embedding.
Voice embedding generated successfully
Voice embedding vector (256-dimensional)
[
-0.021047543734312057,
-0.024001557379961014,
-0.011682840064167976
]Time taken to process the audio and generate the embedding in seconds
13.586525440216064
Model version used to generate the embedding
"titanet-large"
Whether deepfake detection was performed during enrollment
true