OpenAI Realtime API provides conversational speech-to-speech models for developers building interactive voice agents.
OpenAI Realtime API processes audio directly, maintains conversation state, and calls tools during voice interactions. It supports interruptions, natural turn taking, and low-latency spoken responses for conversational applications.
Developers connect browser sessions over WebRTC or server sessions over WebSocket. The Agents SDK provides voice-agent session management, while lower-level interfaces give applications direct control over audio transport, tools, handoffs, and session events.
| Attribute | OpenAI Realtime API | Deepslate Opal | |
|---|---|---|---|
| Provider | |  | |
| Founded | 2015 | 2024 | |
| Sells To | Enterprises, Individuals | Developer | |
| Pricing Model | Subscription, Usage-based | Usage-based | |
| Ownership | Private | Venture-backed |
Provider
Founded 2015
Sells To Enterprises, Individuals
Pricing Model Subscription, Usage-based
Ownership Private
Provider
Founded 2024
Sells To Developer
Pricing Model Usage-based
Ownership Venture-backed
Gemini Live API enables real-time voice and vision interactions through streaming conversational models.
Agent STT is a real-time speech-to-text API package built for AI voice agent workloads.
Gemini Live API enables real-time voice and vision interactions through streaming conversational models.
Agent STT is a real-time speech-to-text API package built for AI voice agent workloads.
Deepgram's Speech-to-Text API is a conversational speech recognition service in real-time and pre-recorded modes.
AssemblyAI's Speech-to-Text API transcribes audio in streaming and pre-recorded modes with speaker diarization.
Developer API for streaming real-time speech-to-speech translation and dubbing over WebSocket.
Amazon Nova Sonic provides streaming speech-to-speech models for conversational applications and voice assistants.
Developer API platform for production-ready speech, transcription, music, agents, and audio generation.
Developer platform providing real-time and batch APIs for speech-to-text, text-to-speech, and voice agents.
Speechmatics' Speech-to-Text API is a multilingual automatic speech recognition service offered in real-time and batch modes with speaker diarization.
Conversational agent platform for configuring, deploying and monitoring voice and chat agents.
Browser-based video conferencing with a built-in AI interpreter for real-time multilingual meetings.
Enterprise AI voice platform for deploying human-like inbound and outbound voice agents at scale.
Besimple AI's conversational audio training-data offering.
Deepgram's Speech-to-Text API is a conversational speech recognition service in real-time and pre-recorded modes.
AssemblyAI's Speech-to-Text API transcribes audio in streaming and pre-recorded modes with speaker diarization.
Developer API for streaming real-time speech-to-speech translation and dubbing over WebSocket.
Amazon Nova Sonic provides streaming speech-to-speech models for conversational applications and voice assistants.
Programmable voice APIs for making, receiving, and managing phone calls globally.
Developer API platform for production-ready speech, transcription, music, agents, and audio generation.
Developer platform providing real-time and batch APIs for speech-to-text, text-to-speech, and voice agents.
Speechmatics' Speech-to-Text API is a multilingual automatic speech recognition service offered in real-time and batch modes with speaker diarization.
Conversational agent platform for configuring, deploying and monitoring voice and chat agents.
Browser-based video conferencing with a built-in AI interpreter for real-time multilingual meetings.
Enterprise AI voice platform for deploying human-like inbound and outbound voice agents at scale.
Besimple AI's conversational audio training-data offering.
