Google added voice remixing to Gemini 3.8 text-to-speech on November 4. Developers can modify an existing prebuilt or designed voice with natural-language instructions through the Gemini API or interactively in Google AI Studio. The API uses gemini-3.8-flash-tts to create a new, independent stored voice and returns a voice ID. Speech can then be synthesized with gemini-3.8-flash-tts or gemini-3.8-flash-lite-tts.
The change connects voice modification with a reusable voice identity: the result is stored separately from the starting voice and can be selected for subsequent speech generation. It also separates the creation step from synthesis, allowing a voice created with Flash TTS to be used with either supported Gemini 3.8 speech model.
