Google AI Studio previews Gemini 3.1 Flash TTS
By Steven Van ·
The text-to-speech model adds inline audio tags, multi-speaker dialogue and 70+ languages, with SynthID watermarking on all output.
Google AI Studio users can now test Gemini 3.1 Flash TTS, a new text-to-speech model rolling out in preview via the Gemini API. The model adds audio tags: natural-language commands embedded in text input that control vocal style, pacing and delivery, alongside native multi-speaker dialogue and support for 70+ languages.
In the AI Studio Playground, developers can set scene direction for multi-turn dialogue, cast speakers with Audio Profiles and Director's Notes for pace, tone and accent, and use inline tags to shift expression mid-sentence. Finished configurations can be exported as Gemini API code to keep voices consistent across projects. All generated audio is watermarked with SynthID.
Read the full announcement.
