Source · release
Voice remixing for Gemini 3.8 TTS
Summary
The Gemini API changelog entry of 2026-11-04 adds voice remixing to Gemini 3.8 TTS: you take a prebuilt or designed voice and change it with natural-language instructions. The result is saved as a new, independent voice that you can use for speech synthesis.
Key points
- Create a remix with
POST /v1beta/voicesusinggemini-3.8-flash-tts, or interactively in Google AI Studio. Either way you get back avoice_...ID. - Use the ID with Gemini 3.8 Flash TTS or Gemini 3.8 Flash-Lite TTS.
- The changelog points to a "Voice remixing" guide for details; pricing and limits were not checked.
Why it matters
Custom voices no longer need audio samples: describing the change is enough, which makes consistent narration voices easy to set up for apps and agents.