Voice Generation
Voice Design returns audition media. An explicit selection creates a persistent provider voice ID. The host keeps each sample’s voice_binding and selects the voice for later speech through the character’s voice.provider_bindings; an uploaded recording alone creates neither.
This operation differs from Speech Synthesis: a voice design establishes how a voice sounds; synthesis speaks supplied text. Both functions currently live in one ElevenLabs package. Voice cloning is not implemented.
Subtypes
Section titled “Subtypes”- Prompt to voice design — implemented
- Save selected design as a provider voice — implemented