OpenAI Realtime API
Low-latency speech-in, speech-out model API
The OpenAI Realtime API streams audio in and out of a speech-capable model over WebSocket, WebRTC or SIP. Developers use it to build low-latency voice assistants without chaining separate speech models.
More in Voice agent platforms
Featured
Vapi
Voice agent platforms
Developer platform for building voice agents
Bland AI
Voice agent platforms
AI agents that place and answer phone calls
Cognigy
Voice agent platforms
Enterprise platform for voice and chat AI agents
Dasha
Voice agent platforms
Conversational AI platform for phone agents
Deepgram Voice Agent API
Voice agent platforms
Speech, language model and voice behind one API
ElevenLabs Agents
Voice agent platforms
Conversational voice agents from ElevenLabs