Voice-first Gemini Live demo: real-time transcription, reasoning and tool calling

AI_Andrew · x · 2026-09-20

Developer AIAndrew built a quick voice-driven app with Google AI Studio and Antigravity, combining Gemini Live with 3.5 Transcribe. The demo does instant live transcription, advanced reasoning and tool calling entirely by voice, inspired by changing autumn leaves. He calls the live transcription "insanely good" while noting the model hedged on whether it's technically still summer. Early exploration, with more to come.

Related event: Developer Builds Open-Source Gemini Live Voice Demo in 20 Minutes(2 posts)→

Original post →

More from Multimodal

Multimodal channel →