{ "@context": "https://schema.org", "@id": "https://aggr.aymericbeaumet.com/items/simonwillison-net/2026-09-15-gemini-live-audio/#webpage", "@type": [ "WebPage", "ArchiveComponent" ], "articleBody": "Google 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\n![Screenshot of a voice chat web interface with a transcript. Top buttons: Start session, End session, Mute mic, plus a Mic level meter and a timer showing 0:33. Status: Listening. Use headphones to reduce echo. Starting a session asks for microphone access. Transcript (with Download transcript and Clear buttons). Gemini: Yes, it's working perfectly. I can hear you clearly. How can I help you today? You: Okay, this is pretty good. Tell me some interesting facts about the California brown pelican. Gemini: They're famous for their spectacular plunge-dives to catch fish and have huge throat pouches that can hold up to three gallons of water and fish. They also made a great recovery after being endangered — marked as Interrupted. You: No, tell me different facts actually. Gemini: They nest in colonies on offshore islands and can hold up to three gallons of water in their throat pouches. At the bottom is a text input reading Or type a message… with a Send button, and the note: Sending a message interrupts the current response. Transcripts may include speech interrupted before playback.](https://static.simonwillison.net/static/2026/gemini-live-tool.webp)\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", "authors": [], "captured_at": "2026-09-16T01:36:55.400042967+00:00", "category": "programming", "content_html": "

Google released Gemini 3.8 Live and 3.8 Live Extended Thinking today - two new speech-to-speech models that are a similar shape to OpenAI's GPT-Live family.

\n

I pointed GPT-6 Astra Extra High at the documentation and had it build me this web UI 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

\"Screenshot

\n

The implementation 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

Here's the Gemini Live tutorial for getting started with that WebSockets API.

\n", "content_markdown": "Google 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\n![Screenshot of a voice chat web interface with a transcript. Top buttons: Start session, End session, Mute mic, plus a Mic level meter and a timer showing 0:33. Status: Listening. Use headphones to reduce echo. Starting a session asks for microphone access. Transcript (with Download transcript and Clear buttons). Gemini: Yes, it's working perfectly. I can hear you clearly. How can I help you today? You: Okay, this is pretty good. Tell me some interesting facts about the California brown pelican. Gemini: They're famous for their spectacular plunge-dives to catch fish and have huge throat pouches that can hold up to three gallons of water and fish. They also made a great recovery after being endangered — marked as Interrupted. You: No, tell me different facts actually. Gemini: They nest in colonies on offshore islands and can hold up to three gallons of water in their throat pouches. At the bottom is a text input reading Or type a message… with a Send button, and the note: Sending a message interrupts the current response. Transcripts may include speech interrupted before playback.](https://static.simonwillison.net/static/2026/gemini-live-tool.webp)\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", "dateCreated": "2026-09-16T01:36:55.400042967+00:00", "date_created": "2026-09-15T22:47:07+00:00", "date_modified": "2026-09-15T22:47:07+00:00", "date_published": "2026-09-15T22:47:07+00:00", "description": "Tool: Gemini Live audio Google released Gemini 3.8 Live and 3.8 Live Extended Thinking today - two new speech-to-speech models that are a similar shape to OpenAI's GPT-Live family. I pointed GPT-6 Astra Extra High at the documentation and…", "external_url": "https://simonwillison.net/2026/Sep/15/gemini-live/", "git": { "commit": "ebb0941c5c06a2ab2377342d607535262a3e969e", "path": "items/simonwillison-net/2026/09/2026-09-15-gemini-live-audio.md", "permalink": "https://github.com/aymericbeaumet/aggr-instance/blob/ebb0941c5c06a2ab2377342d607535262a3e969e/items/simonwillison-net/2026/09/2026-09-15-gemini-live-audio.md" }, "id": "urn:aggr:item:07fce4b522e185202bb241f852ddf6c180094873", "isBasedOn": { "@id": "https://simonwillison.net/2026/Sep/15/gemini-live/" }, "mainEntity": { "@id": "https://simonwillison.net/2026/Sep/15/gemini-live/", "@type": "CreativeWork", "archivedAt": { "@id": "https://aggr.aymericbeaumet.com/items/simonwillison-net/2026-09-15-gemini-live-audio/#webpage" }, "dateModified": "2026-09-15T22:47:07+00:00", "datePublished": "2026-09-15T22:47:07+00:00", "headline": "Gemini Live audio", "timeRequired": "PT1M", "url": "https://simonwillison.net/2026/Sep/15/gemini-live/", "wordCount": 124 }, "name": "Archived snapshot: Gemini Live audio", "preview": null, "reading_minutes": 1, "replicated_at": null, "source": "simonwillison-net", "source_name": "Simon Willison's Weblog", "summary": "Tool: Gemini Live audio Google released Gemini 3.8 Live and 3.8 Live Extended Thinking today - two new speech-to-speech models that are a similar shape to OpenAI's GPT-Live family. I pointed GPT-6 Astra Extra High at the documentation and had it build me this web UI 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. The implementation 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. Here's the Gemini Live tutorial for getting started with that WebSockets API. Tags: google, tools, websockets, generative-ai, llms, gemini, llm-release, speech-to-text", "tags": [ "gemini", "generative-ai", "google", "llm-release", "llms", "speech-to-text", "tools", "websockets" ], "temporalCoverage": "2026-09-16T01:36:55.400042967+00:00", "title": "Gemini Live audio", "url": "https://aggr.aymericbeaumet.com/items/simonwillison-net/2026-09-15-gemini-live-audio/", "word_count": 124 }