- What it is: ChatGPT voice mode lets you talk to ChatGPT out loud and hear a spoken reply, mostly in the mobile app.
- How to start: Open the ChatGPT app, tap the voice icon next to the message box, and start speaking.
- By plan (verified on chatgpt.com/pricing, 11 Sep 2026): Free gets limited voice chats; Go and Plus get expanded voice; Pro is unlimited within abuse guardrails.
- Best for: language practice, hands-free cooking, rehearsing an interview, or just thinking out loud.
ChatGPT voice mode is the feature that lets you speak to ChatGPT and hear it reply, instead of typing. You open the ChatGPT mobile app, tap the voice icon next to the message box, and start talking. It works on the free plan with some limits, and gets more generous on the paid plans. That’s the short answer — the rest of this guide covers how to use it well, what each plan actually gives you, and where voice really pays off.
How to Start a Voice Conversation
Voice mode lives in the official ChatGPT app on iPhone and Android. Once you’ve signed in:
- Open a new chat (or continue an existing one).
- Look at the bar where you’d normally type. Next to it there’s a small voice icon.
- Tap it. Give the app permission to use your microphone the first time.
- Speak naturally. ChatGPT will listen, then reply out loud in a synthetic voice.
- To end the conversation, tap the stop or close control on the voice screen.
You can interrupt mid-sentence — if ChatGPT is rambling, just start speaking and it will stop and listen. That’s what makes it feel like a real conversation rather than a walkie-talkie.
On desktop (Windows and macOS), the ChatGPT desktop app also supports spoken conversations. The exact button placement changes with app updates, but the pattern is the same: look for a headphones or waveform icon near the message box. If you don’t see it, check for an app update in the Microsoft Store (Windows) or Mac App Store.
A quick note on terms. A model is the AI brain behind ChatGPT. Tokens are the small chunks of text (roughly 3–4 characters each) the model reads and writes. The context window is how much text the model can hold in mind during one conversation. You don’t need to think about tokens while chatting — but they matter when we get to plan limits below.
What Each ChatGPT Plan Gives You for Voice
Voice access varies by plan. These figures were verified on chatgpt.com/pricing on 11 September 2026.
| Plan | Voice access | Fast-model context | Reasoning context |
|---|---|---|---|
| Free | Limited voice chats | 27K tokens | — |
| Go | Expanded voice | 54K tokens | 256K tokens |
| Plus | Expanded voice | 54K tokens | 256K tokens |
| Pro | Unlimited (within abuse guardrails) | 128K tokens | 400K tokens |
Those context numbers are the ChatGPT app limits, not the developer API. If you’ve read that a model like GPT-6 Astra has a roughly 1.05 million token context window, that’s the API figure — different product, different limits. Inside the ChatGPT app, what you get is the number in the table above.
If you’re weighing up which plan to pay for, our full breakdown of ChatGPT Free vs Go vs Plus vs Pro compares everything side by side, not just voice.
Everyday Situations Where Voice Beats Typing
Voice mode isn’t just a gimmick. There are moments when it’s genuinely better than tapping on a screen.
Practising a Language
You can tell ChatGPT: “Let’s have a five-minute conversation in Spanish about ordering food in a restaurant. Correct me gently if I make mistakes.” It will play the waiter, you play the customer, and you get real speaking practice without the anxiety of a real person judging you. Ask it to switch to English for a moment if you get stuck.
Hands-Free Cooking
Flour on your hands, phone on the counter. “What temperature for a Victoria sponge, and how long?” Follow-up: “Convert that to Celsius.” ChatGPT can walk you through a recipe step by step without you having to touch the screen. Pair it with a smart speaker or Bluetooth kitchen speaker for better sound over the extractor fan.
Rehearsing a Job Interview
Tell it the role you’re applying for and paste (or read) the job description. Ask it to interview you as if it were the hiring manager, one question at a time, and give feedback after each answer. Speaking your answers out loud is very different from typing them — it surfaces the “ums” and the rambly bits you’d otherwise not notice.
Talking Through a Decision
Sometimes you just need to think out loud with something that will actually listen and ask reasonable follow-up questions. Choosing between two flats, weighing up a career move, planning a difficult conversation — voice mode is surprisingly good as a sounding board because it forces you to articulate things you’d normally keep vague in your head.
Accessibility
For anyone with dyslexia, low vision, arthritis, or just a long day of typing at work, being able to speak to a computer and hear a reply is a real quality-of-life upgrade — not a novelty.
Privacy: What to Know Before You Start Talking
Voice conversations are still conversations. Your speech is sent to OpenAI’s servers, converted to text, processed, and a reply is generated and spoken back. That means the usual rules apply: don’t say things you wouldn’t want stored, and don’t share other people’s private information (medical details, bank data, passwords) unless you’re comfortable with that.
You can control whether your chats are used to improve OpenAI’s models, and you can delete conversations from your history. The exact toggles move around between app updates, but they live under Settings → Data Controls. For the current official policy, see OpenAI’s privacy policy and the data controls page linked from within the app.
One more thing to be aware of: the AI voice sounds convincing, but it can still be wrong. It will happily state a wrong date, a made-up statistic, or a bad recipe conversion in a confident tone. Voice makes errors feel more authoritative because we’re wired to trust a fluent human-sounding speaker. Sanity-check anything that matters — medicine dosages, legal advice, tax numbers — against a proper source.
The Free Alternative: Gemini Live
If you’re on the free ChatGPT plan and hitting the voice limits, Google’s equivalent is worth knowing about. Gemini Live — Google’s spoken conversation mode — is included in Gemini’s free plan (verified on gemini.google/subscriptions on 11 September 2026). You get it inside the Gemini app on iPhone and Android.
The experience is similar: tap a voice icon, talk naturally, hear a reply. The personality is a little different — some people find Gemini Live warmer, others prefer ChatGPT. Try both and see which one you’d actually keep talking to. Our full write-up of what Gemini’s free plan includes covers what else you get for nothing.
For more options beyond the two big names, our roundup of ChatGPT alternatives lists the assistants worth trying, and if you’re mainly comparing writing style and reasoning, Claude vs ChatGPT is a good next read.
Tips for Better Voice Conversations
- Set the scene first. A one-line instruction (“Be my Spanish tutor and only reply in Spanish”) at the start of the chat shapes everything that follows.
- Speak in natural chunks, not one word at a time. The model does better with a whole sentence than with hesitant fragments.
- Interrupt when it goes off-track. You don’t have to wait for it to finish. Just start talking.
- Use headphones in public. The synthesised voice is loud enough on speaker to embarrass you on a train.
- Fall back to text for anything precise. Numbers, code, names of medications — type or read those back to check, don’t rely on the spoken transcript.
Frequently Asked Questions
Is ChatGPT voice mode free?
Yes, there’s a limited version on the free plan. If you hit the cap, you either wait for it to reset or upgrade to Go, Plus, or Pro. Pro is unlimited within OpenAI’s abuse guardrails (verified on chatgpt.com/pricing on 11 September 2026).
Can I use ChatGPT voice mode on a laptop?
Yes. The ChatGPT desktop app for Windows and macOS supports spoken conversations. Look for the voice icon near the message box in the app. If it isn’t there, update the app — the interface changes between releases.
Does voice mode work offline?
No. ChatGPT voice mode needs an internet connection because your speech is processed on OpenAI’s servers, not on your phone. If your connection drops, the conversation pauses.
Can ChatGPT copy my voice?
The official ChatGPT app does not clone your voice — you speak, and it replies in one of its own preset voices. Voice cloning is a different category of tool entirely; see our roundup of AI voice cloning tools if that’s what you need.
Which languages does voice mode support?
ChatGPT voice mode understands and speaks many major languages, and it can switch mid-conversation if you ask. Coverage and accent quality vary by language and improve over time. For anything long or important, cross-check with a dedicated translation tool.
Is what I say to ChatGPT private?
Your voice is sent to OpenAI’s servers to be processed. You can control whether it’s used to train future models and you can delete chats from your history, under Settings → Data Controls in the app. Don’t say anything to it that you wouldn’t want stored on a company’s servers.