Voices
A voice from a few seconds of audio
Clone from a short reference clip, or design one from age, accent and delivery. Sixteen synthesis engines behind one interface, so the model can change without the workflow changing.
- Clone
- Record or upload a few seconds of clean speech. Zero-shot — no training run, no waiting.
- Design
- No reference clip? Describe the voice instead: age, accent, pitch, delivery.
- Direct
- Per-line instructions change how a line is read, not just what it says.
- Own it
- Mark a profile as your own voice with a recorded consent phrase. That flag is what lets an agent use it on a call.
Everything the app synthesizes carries an inaudible AudioSeal watermark, so audio it made stays identifiable as audio it made. On the desktop app it is a setting; for agents it is not.