LiveKit says its Inference service now runs three voice-agent architectures: cascaded STT-LLM-TTS, turn-based speech-to-speech, and full-duplex models that listen and talk simultaneously.
It lists GPT-Live and GPT Realtime from OpenAI, Voice Think Fast from xAI, and Gemini Live from Google DeepMind as supported models. Developers can access all of them with one LiveKit Cloud key and one bill.
Voice agents get built three ways: cascaded (STT, LLM, TTS), turn-based speech-to-speech, and full-duplex, where the model listens and talks at the same time.
All three now run on LiveKit Inference: GPT-Live and GPT Realtime from @OpenAI, @grok Voice Think Fast from @xai, and Gemini Live from @GoogleDeepMind. One LiveKit Cloud key, one bill.
Source: LiveKit · x.comPublished · added here
