LiveKit Inference adds voice models from OpenAI, xAI and Google DeepMind

LiveKit says its Inference service now runs GPT-Live and GPT Realtime from OpenAI, Voice Think Fast from xAI, and Gemini Live from Google DeepMind.

LiveKit describes the models as covering three voice-agent architectures: cascaded speech-to-text, LLM and text-to-speech; turn-based speech-to-speech; and full-duplex models that listen and talk at the same time.

According to LiveKit, developers can use all of these models with one LiveKit Cloud key and one bill.