Voices and voice cloning

Pick the voice your agent speaks in, or clone one from a consented recording.

Enlarge

1. Open Voices and play samples in any language. Samples are free.
2. To use a voice, open the agent, click Edit and pick it under the voice for calls. For the voice API, copy the voice's ID.

Clone a voice

1. Go to My voices and click Clone a voice.
2. The speaker reads one consent sentence aloud (about 10 seconds), or you upload a recording. The recording proves they agreed and becomes the voice.
3. The voice is private to your workspace: use it for your agents and in the voice API with the tts-clone model.

Related guides