OpenAI Realtime API

API for real-time multimodal communication with models over WebRTC, WebSocket, and SIP, supporting speech-to-speech, text, image, and audio inputs and outputs with ultra-low latency.