Gemini 3.8 TTS Playground
Google just dropped Gemini 3.8 TTS with 2,000+ voices and a 30‑second custom‑voice option, and I built a playground to test it.
Why Now
Google released two new Gemini text‑to‑speech models, gemini‑3.8‑flash‑tts and gemini‑3.8‑flash‑lite‑tts, today.
What Happened
The models support over 2,000 voices and allow custom voices from a 30‑second audio sample. The playground lets users compose single‑voice or multi‑speaker conversations, preview audio, and view request/response details. Generating 1m 18s of audio with Gemini 3.8 Flash TTS cost 2.74 cents and took about 20 seconds.
Why It Matters
The large voice library and easy custom‑voice creation lower barriers for developers to build rich, personalized audio experiences. Multi‑speaker support enables realistic dialogue simulations, useful for virtual assistants, games, and accessibility tools. The cost and speed metrics suggest it’s practical for moderate‑scale production use.
The Limitation
The playground relies on an open CORS policy and a GPT‑6 Astra backend; results may vary with different API keys or network conditions.
What You Can Do
Try the playground with your own Gemini API key to prototype a custom‑voice narration or dialogue.
Source
Read original sourceWhy we picked this
New Gemini TTS models and playground tool release.