Google's voice-design walkthrough goes from text prompts to dialogue
Thor Schaeff demonstrates custom narration and voice replication as Google for Developers has an unusually busy day.
Why this story? The whole channel is running at 2,256,660 views a day, 12 times a normal day (187,951).
A voice described in words is the starting point for Google for Developers' new speech-generation walkthrough. Google DeepMind's Thor Schaeff presents tools for designing a custom voice from a text prompt, replicating your own voice and putting multiple speakers into a generated conversation.
"Design and replicate voices with Gemini 3.8 Flash TTS" is Google's own product tutorial. It moves from Google AI Studio into code, covering how to use custom voices across languages, manage them through the Gemini API and generate multi-speaker dialogue. A live podcast demo comes toward the end, giving the technical instructions a more recognizable destination: a conversation rather than an isolated line of narration.
The upload arrives alongside a sharp channel-wide rise: roughly 2.3 million views in a day, about twelve times the usual pace. That jump cannot be pinned on this tutorial, whose audience is much smaller.
Voice experiments are a recurring subject in the channel's October uploads. Sam Goldstein presents custom story narration in "📖 Explore custom voice narration with Gemini 3.8 Flash TTS," while a separate demonstration by Matias Piña Aguilera uses laptop movement to control browser-based music.
Published on Trari
Where next?
- Channel
- Google for Developers
- Subscribers
- 2,680,000
- Channel views, last 7 days
- 4,025,720
- Channel views, pace per day
- 2,256,660 (on a normal day: 187,951)
The numbers
Channel-wide views have risen sharply. The voice-design tutorial has a much smaller audience than that daily total.
💡 Test your knowledge
Which Google tool does Schaeff use for the voice-design walkthrough?
The tutorial demonstrates voice design and voice replication in Google AI Studio.



