Voice Design v3
Voice Design v3 creates candidate voices from a written description. The June 2025 release generates three previews, then lets the user save a chosen voice for speech generation.
Design and remix
The API supports reference audio and a control that balances it against the prompt. Voice Remixing uses this same model. A remix creates a new voice profile without changing the original.ElevenLabs Voice Design APIElevenLabs Voice Remixing capabilities
Practical guidance. Save the chosen voice ID before generating a series of scenes. Repeating the description for each scene can create different candidates rather than preserve one character.ElevenLabs: Voice Design v3 release
Specifications & access
| Detail | Value and conditions |
|---|---|
| Launch date | |
| Access | API, Provider product |
| Model identifiers | eleven_ttv_v3 |
| API identifier | eleven_ttv_v3ElevenLabs model catalog and deprecations |
Related models
- Eleven v3
ElevenLabs · Eleven speech
Expressive speech model with audio tags and multi-speaker performance.
Launch date API · Provider product - Multilingual Voice Design v2
ElevenLabs · Eleven Voice Design
Multilingual voice-design model for generating voices from descriptions.
Launch date not confirmedAPI