{"slug": "gemini-live-audio", "title": "Gemini Live audio", "summary": "Google released Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking, two new speech-to-speech models that mirror the shape of OpenAI's GPT-Live family. Developer Simon Willison used GPT-6 Astra Extra High to build a browser-based web UI for the models that connects to the wss://generativelanguage.googleapis.com WebSocket endpoint and uses the Web Audio API AudioContext for capture and playback, letting users select a model and voice preset, enter an optional system prompt, and interrupt the model mid-speech.", "body_md": "**Tool:** [Gemini Live audio](https://tools.simonwillison.net/gemini-live)\n\nGoogle released [Gemini 3.8 Live and 3.8 Live Extended Thinking](https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-8-live-gemini-3-8-live-extended-thinking/) today - two new speech-to-speech models that are a similar shape to OpenAI's [GPT-Live](https://openai.com/index/introducing-gpt-live/) family.\n\nI pointed GPT-6 Astra Extra High at the documentation and [had it build me this web UI](https://gist.github.com/simonw/067b7430c5b1f743af9419b0184c38ef) for trying out the new models. You can select a model and voice preset, enter an optional system prompt and then start a voice conversation through your browser, including the ability to interrupt the model while it is talking.\n\nThe [implementation](https://github.com/simonw/tools/blob/main/gemini-live.html) uses no libraries. It connects to the `wss://generativelanguage.googleapis.com/ws/google.ai.generativelanguage.v1alpha.GenerativeService.BidiGenerateContent?key=...` WebSocket endpoint and uses a Web Audio API `AudioContext` for both capture and playback.\n\nHere's [the Gemini Live tutorial](https://ai.google.dev/gemini-api/docs/live-api/get-started-websocket) for getting started with that WebSockets API.\n\nTags: [google](https://simonwillison.net/tags/google), [tools](https://simonwillison.net/tags/tools), [websockets](https://simonwillison.net/tags/websockets), [generative-ai](https://simonwillison.net/tags/generative-ai), [llms](https://simonwillison.net/tags/llms), [gemini](https://simonwillison.net/tags/gemini), [llm-release](https://simonwillison.net/tags/llm-release), [speech-to-text](https://simonwillison.net/tags/speech-to-text)", "url": "https://wpnews.pro/news/gemini-live-audio", "canonical_source": "https://simonwillison.net/2026/Sep/15/gemini-live/", "published_at": "2026-09-15 22:47:07+00:00", "updated_at": "2026-09-16 16:11:12.522542+00:00", "lang": "en", "topics": ["generative-ai", "large-language-models", "ai-products", "ai-tools", "natural-language-processing"], "entities": ["Google", "Gemini 3.8 Live", "Gemini 3.8 Live Extended Thinking", "OpenAI", "GPT-Live", "Simon Willison", "GPT-6 Astra Extra High", "Web Audio API"], "alternates": {"html": "https://wpnews.pro/news/gemini-live-audio", "markdown": "https://wpnews.pro/news/gemini-live-audio.md", "text": "https://wpnews.pro/news/gemini-live-audio.txt", "jsonld": "https://wpnews.pro/news/gemini-live-audio.jsonld"}}