Guide · GPT-live-1 voice

Talk to GPT-live-1 in your browser

GPT-live-1 is OpenAI's real-time voice model for natural spoken conversation. This workspace connects your microphone to the official GPT-live-1 API over WebRTC, so you can talk, interrupt and be interrupted like a phone call.

Start a voice session

In short

Full-duplex: it can listen while it speaks, so interruptions feel natural.
WebRTC connects your microphone directly to the live session.
1 credit reserves 1 minute; unused whole minutes are returned if you end early.
No free trial; prepaid minute packs, no subscription.

Getting started

Choose how many minutes to reserve as your spending ceiling.
Allow microphone access when the browser asks.
Talk naturally — you can interrupt while GPT-live-1 is speaking.
End the session and unused whole minutes return to your balance.

Why it feels different from voice chat apps

Many voice assistants record, transcribe, answer in text and then read the answer aloud. GPT-live-1 works on live audio in both directions, which reduces the pause before replies and lets you cut in mid-sentence.

Privacy and keys

The OpenAI API key stays on the server; the browser only receives the WebRTC answer. The workspace does not upload a recording to its own media library. Avoid discussing confidential information you are not permitted to process.

Frequently asked questions

Is this the official OpenAI site?

No. It is an independent web app that connects to the official GPT-live-1 API.

What happens when my reserved time runs out?

The session ends at the limit so it cannot keep consuming paid time.

Is "GPT 1 live" the same thing?

People search for it as GPT 1 live, GPT live 1 or GPT-live-1; this workspace uses the gpt-live-1 model.

Updated 2026-10-04. GPT-live-1.pro is an independent service using the official OpenAI API. Voice is AI-generated; verify important information before relying on it.