30 seconds: that's how Gemini 3.8 clones your voice
Google launched Gemini 3.8 Flash TTS with voice replication: Gemini clones your voice from a short sample, with verified consent, in over 100 languages.
On September 24, 2026, Google launched Gemini 3.8 Flash TTS and Flash-Lite TTS, its most expressive voice models yet. The feature drawing the most attention is that Gemini clones your voice from a short audio sample, just 30 seconds long.
How Gemini clones your voice
The process is called voice replication: you upload roughly 30 seconds of audio, and the model generates a voice profile from it. Before creating that profile, Gemini requires consent verification — a step meant to stop anyone from cloning someone else's voice without their explicit authorization.
Over 100 languages, including Mexican Spanish
The cloned voice isn't locked to a single language: the system supports more than 100 languages and dialects, including Mexican Spanish. It also lets you direct the performance line by line — pace, accent, and even pauses — instead of producing one flat audio track from start to finish.
The honest caveat: it's an API, not a button
This is a capability of the Gemini API, built for creators and developers, not a button inside the consumer Gemini app. And as with any voice cloning, verified consent is the real line between a useful production tool and an impersonation risk.
How to apply it in your business?
If your team regularly produces video or audio — training material, ads, content in multiple languages — recording 30 seconds once and reusing that cloned voice can replace the need to hire a voice actor for every new project. Before rolling it out, it's worth defining internally who can authorize cloning a voice (your own or a teammate's) and keeping that consent documented, since that's exactly the step Gemini requires before generating the profile.
That single change turns a one-off actor booking into a reusable asset the team controls directly.
For a team producing weekly episodes or multilingual training material, that difference adds up fast across dozens of videos.
Conclusion
The fact that Gemini clones your voice in 30 seconds doesn't mean anyone can do it without permission: verified consent remains the filter. But for anyone producing content in multiple languages or needing a consistent voice across videos, this API capability changes the real cost of production.