GPT-Live (gpt-live-1): output audio frames generated and recorded, but not delivered to the client (WebSocket and WebRTC)

🌐 链接: https://community.openai.com/t/gpt-live-gpt-live-1-output-audio-frames-generated-and-recorded-but-not-delivered-to-the-client-websocket-and-webrtc/1410537
🏷️ 分组: OpenAI Developer Community
🕒 时间: 2026-10-29 03:28:00
OpenAI Developer Community
GPT-Live (gpt-live-1): output audio frames generated and recorded, but not delivered to the client (WebSocket and WebRTC)
GPT-Live (gpt-live-1): output audio frames generated and recorded, but not delivered to the client (WebSocket and WebRTC)
Environment Surface: API Endpoint or feature: Live sessions (/v1/live/sessions), WebSocket and WebRTC transports Model: gpt-live-1, delegation backend gpt-5.6-luna SDK: livekit-agents 1.8.9 (livekit-plugins-openai GPTLiveModel), Python 3.12 Summary On the…