Create a custom voice (clone)
Audio
Create Voice
Clone a custom voice from a short reference sample (a few seconds of speech). model selects the provider to clone on (e.g. mistral/voxtral-mini-tts-2603) — the returned voice works with any of that provider’s TTS models; audio is the reference sample as a file_<id>, an https URL, or a base64 data-URI; languages and gender are optional hints. Returns an opaque id (voice_…) to pass as voice on POST /v3/audio/speech. The voice is scoped to your project — it is not visible to or usable by other projects — and the reference audio is not stored. Paid feature.
POST
Create a custom voice (clone)
Authorizations
API key authentication. Pass your API key as a Bearer token.
Body
application/json