Design and replicate voices with Gemini 3.8 Flash TTS
Gemini TTS now replicates any voice from 20 seconds of audio and designs new voices from text descriptions
“So that was my voice that I just replicated here in Google AI Studio from just 20 seconds of audio.”
Google unveiled voice design and voice replication as major upgrades to Gemini TTS, demonstrated live at Google I/O. Voice design lets developers generate custom voices from natural-language descriptions (e.g., accent, age, energy level), while voice replication clones a speaker from just 20 seconds of audio. The capabilities extend to multi-speaker dialog generation and are fully accessible via the Gemini API, lowering the barrier to production-quality synthetic speech significantly.