ChatGPT Voice Mode: How to Use It and What to Know

Some of my best ideas arrive while I'm walking, not while I'm sitting at a keyboard. Talking to ChatGPT out loud is the closest thing I've found to thinking with a sharp colleague who never gets tired of the same question. The catch is that the settings changed a lot, so a lot of the advice online is out of date.

This guide follows OpenAI's current help page for ChatGPT Voice. It covers what the three voice options are, how to start a conversation, what it can and can't do, and what happens to your recordings. I've added five prompts for the moments when typing isn't practical.

What is ChatGPT voice mode?

ChatGPT Voice lets you speak with ChatGPT and hear a spoken answer. It works inside a normal chat, so you can follow the answer as text while you listen, type when you can't speak, and scroll back through earlier messages without starting again.

OpenAI also offers Dictation, which is different. Dictation turns one recording into editable text that you review and send. Voice is for a live back-and-forth conversation.

What are the Live, Advanced and Standard voice options?

Under Settings, then Voice, you may see three options, depending on your plan, workspace, region and app version:

• Live: listens and speaks at the same time, so you can interrupt or carry on talking while it answers. It can use web search and memory, show visual results, and work with text and images in the same chat. It can also use the plugins and connected apps available to your account. Live doesn't support video or screen sharing.
• Advanced: the earlier real-time voice experience. Use this one when you need video or screen sharing on mobile.
• Standard: a turn-by-turn style that transcribes what you say before it replies.

To switch, open Settings, choose Voice and pick Live, Advanced or Standard.

How do I start a voice conversation?

On the web, go to ChatGPT.com and select the Voice icon in the prompt window. On iPhone or Android, select the Voice icon in the message bar. If you're asked, allow microphone access, and the first time choose a voice. Then just start talking.

During the conversation, the microphone control mutes and unmutes you, and the exit control ends it. You can also attach an image with the add button or type a message while Voice is active, and ChatGPT will answer without starting a new chat.

How do I change the voice or the language?

Open Settings, then Voice, then Voice to choose from nine options, each with its own personality: Arbor is easygoing and versatile, Breeze animated and earnest, Cove composed and direct, Ember confident and optimistic, Juniper open and upbeat, Maple cheerful and candid, Sol savvy and relaxed, Spruce calm and affirming, and Vale bright and inquisitive.

For language, choose Language in the same place. Picking the language you speak most helps ChatGPT understand you, and you can also ask it mid-conversation to speak another language. You can ask it to talk faster, slower or in a different tone, although exact playback speed controls aren't available and preset personalities don't apply to Live.

How can I use voice mode for my business?

These are the jobs where speaking beats typing for me. Copy any of them and swap in your details.

Copy this prompt

I'm walking to a meeting with [client or supplier] in ten minutes.
Ask me three questions to find out what I want from it, then give me a spoken summary of my goals and one opening line.

Copy and paste this prompt

Let me talk through an idea for [offer or product].
Don't interrupt until I say I'm finished. Then tell me what's unclear, what's strong and the first thing I should test.

Copy this prompt

I'm going to describe a job I did for a customer.
Turn it into a short case study with a problem, what I did and the result, in plain language. Then read it back to me.

Copy and paste this prompt

Role-play a difficult customer who says [objection].
Respond as they would. After three exchanges, stop and tell me how I could have answered better.

Work with me

Want AI doing the heavy lifting in your marketing?

I build the systems that handle the boring 80 percent, so you get your week back. Done properly, with the human kept in.

Book a free 30-minute callOr grab the free resources →

Copy this prompt

Quiz me on [topic I'm learning], one question at a time.
Wait for my answer, tell me if I'm right and explain briefly before moving on.

Two small tricks help. At the start, you can ask Live to wait until you're ready, for example, "Wait until I ask you to respond." And if you want to dictate something to polish later, switch to Dictation instead, which gives you text you can edit.

Can I keep talking while I use other apps or drive?

You can keep a conversation going while you use other apps or when your phone is locked by switching on Background conversations under Settings, then Voice. It ends when you end it, force close the app, reach a usage limit or hit the maximum session length.

ChatGPT also works in Apple CarPlay on supported iPhones, where you can start a voice conversation or continue a recent chat. OpenAI's own reminder applies: only use your device when the law and conditions allow, set things up before you drive, and don't interact with it while the vehicle is moving.

Can voice mode handle meetings with several people?

Not well yet. OpenAI says Live is built mainly for one-to-one conversation. It copes with background noise, but it isn't yet optimised for several speakers and may reply when people are talking to each other rather than to it. Headphones and a quiet room help, and on iPhone you can turn on Voice Isolation from Control Center under Mic Mode.

What happens to my voice recordings?

This is the part to read before you talk about a client. OpenAI says audio clips from Live and Advanced conversations are stored alongside the transcript in your chat history and kept for 30 days. If you delete the chat, its clips are deleted within 30 days too, unless they must be kept for security, safety or legal reasons.

With Standard, your audio is deleted once it has been transcribed. OpenAI says it doesn't train its models on your audio or video clips unless you choose to share them. If Improve the model for everyone is on, transcripts and other files may still be used, depending on your plan and settings.

The transcript isn't a word-for-word record either, so don't treat it as proof of what anyone said. I've written more on this in my guide to whether ChatGPT is private.

Tools that work with voice mode

Voice is at its best next to the apps you already use. Live can use the plugins and connected apps available to your account, and on web and mobile it asks you to approve actions on screen rather than by voice. A voice recorder app, Apple Notes or Google Keep are handy places to park what you decide, and a simple project setup keeps the follow-up organised: see how ChatGPT projects work.

If you'd like help turning these habits into a routine for your team, this is how I work with business owners.

I keep every related walkthrough in the ChatGPT Prompts and Limits: 30 Guides to Fun Prompts, Photo Edits and Daily Caps.

Frequently asked questions

What is ChatGPT voice mode?

It lets you speak with ChatGPT and hear spoken answers inside a normal chat, so you can follow along in text. OpenAI offers three options in Settings under Voice: Live, Advanced and Standard.

Can I interrupt ChatGPT while it's talking?

Yes. With Live, ChatGPT can listen and speak at the same time, so you can interrupt or carry on. Noise, overlapping speech and microphone settings can still affect what it hears.

Does ChatGPT voice mode record me?

OpenAI says audio clips from Live and Advanced conversations are stored with the transcript and kept for 30 days. Standard deletes the audio after transcription.

Can ChatGPT voice mode share my screen or use my camera?

Live doesn't support video or screen sharing. OpenAI says those stay available to eligible subscribers on the iPhone and Android apps when you use Advanced.

What is the difference between voice mode and dictation?

Voice is a live conversation. Dictation records a prompt, shows you an editable transcript and sends it as text when you're ready.