For the complete documentation index, see llms.txt. Markdown versions of documentation pages are available by appending .md to the page URL.
Primary navigation

Audio and voice

Start with GPT-Live for conversational voice, or choose an API for another audio task.

For a new conversational voice application, start with GPT-Live. It can listen while speaking and keep the conversation moving while a backend agent reasons, uses tools, or completes a task.

Connect your first conversation with the WebRTC quickstart, then write a short Live prompt. If you already have a Realtime application or text agent, follow Migrate to GPT-Live.

Choose another audio workflow

Use the Realtime API when you need its session and tool model. For transcription, translation, or speech generation without a conversational agent, choose the dedicated API below.

BuildStart hereWhat you control
A speech-to-speech agent using the Realtime session and tool modelRealtime APIAudio turns, session state, tools, and interruptions.
A voice interface for an existing text agentVoice agentsSpeech-to-text, the text-agent workflow, then text-to-speech.
A transcript of an audio fileFile transcriptionFile uploads, bounded requests, and supported transcript formats.
Live captions without assistant speechLive transcriptionStreaming audio and incremental transcript events.
Continuous speech translationLive translationA dedicated translation session, not a voice-agent turn loop.
Narration or generated speechText to speechText, voice, and output format.
Audio input or output in an existing chat appAudio in Chat CompletionsBounded multimodal chat requests.

Build with voice

Use Voice agents to compare architectures. Start with the prompting guide for GPT-Live or Realtime. Then use the shared guides for custom voices, evaluation, and cost optimization. Each guide distinguishes model- or API-specific behavior.

Choose a connection

For browser audio, start with WebRTC. For server audio pipelines, use WebSockets. For phone calls, see Telephony and SIP. A server-side control connection lets a trusted backend observe and control a media session.

Select your API on each connection page. Sharing a transport does not make GPT-Live and Realtime handshakes, credentials, or event formats interchangeable. Check the connection guide for prerequisites and setup instructions.

Add audio to your existing application

The Chat Completions examples now live in Audio in Chat Completions. For a browser voice-agent starter, use the GPT-Live WebRTC quickstart.