Skip to content

ChatGPT Voice: steering work by talking

Where ChatGPT Voice is available, how to start a voice chat, how it differs from voice dictation, using it to coordinate several tasks, and the care that macOS screen context deserves.

CurrentLast verified

Platforms

  • ChatGPT desktop app (Chat / Work / Codex)
  • Remote on iOS (after pairing a desktop host)

What the official documentation says

  • ChatGPT Voice is powered by GPT-Live and lets you talk through ideas and coordinate tasks in Chat, Work and Codex in the ChatGPT desktop app.

    ChatGPT Voice

  • ChatGPT Voice is available in the desktop app with ChatGPT Plus, Pro, Business, Edu and Enterprise plans; Enterprise and Edu availability begins with a two-week early-access period before the feature becomes available by default.

    ChatGPT Voice

  • You can also use ChatGPT Voice through Remote on iOS after pairing your phone with a desktop host.

    ChatGPT Voice

  • You start with Start voice chat in an existing task, or Start new voice chat in a new chat or task; the first time you also allow microphone access, choose a voice, and review screen context on macOS.

    ChatGPT Voice

  • A shortcut can be set in Settings > Voice > Voice chat hotkey.

    ChatGPT Voice

  • ChatGPT Voice supports natural turn-taking — you can interrupt ChatGPT during a response, ask a follow-up, or change direction.

    ChatGPT Voice

  • ChatGPT Voice can start separate tasks for longer work, check existing tasks and send follow-up instructions, and brings progress, blockers and results back to the voice conversation.

    ChatGPT Voice

  • You can ask to speak to another Codex task and then return to the previous one; the destination task must support voice and be available on a connected host.

    ChatGPT Voice

  • ChatGPT Voice follows the same permissions as the tasks it directs.

    ChatGPT Voice

  • On macOS, turning on Screen context in Settings > Voice lets ChatGPT take an appshot of your frontmost window and use it as context; your organization can disable this capability.

    ChatGPT Voice

  • An appshot can include the window's image and accessible text, including content outside the visible scroll area, and the documentation advises avoiding windows that contain sensitive information.

    ChatGPT Voice

  • Only one voice chat can be active across the ChatGPT desktop app at a time, and voice conversations use a separate, plan-dependent allowance measured in rolling five-hour windows.

    ChatGPT Voice

  • Tasks started through Voice continue to use your Codex usage budget.

    ChatGPT Voice

It is for steering, not for reading aloud

ChatGPT Voice is powered by GPT-Live and lets you talk through ideas and coordinate tasks in Chat, Work and Codex in the ChatGPT desktop app. The documented framing: start work, check progress, or change direction without switching back to typing.

The valuable phrase there is change direction. Voice is not primarily a faster input method — it is a way to intervene while something is already happening.

Availability

ChatGPT Voice is available in the ChatGPT desktop app with Plus, Pro, Business, Edu and Enterprise plans. Enterprise and Edu availability begins with a two-week early-access period before the feature becomes available by default.

You can also use it through Remote on iOS after pairing your phone with a desktop host.

Availability also depends on rollout status and workspace settings.

Starting a voice chat

  1. Open the Codex task you want to discuss in the desktop app, or start a new chat or task.
  2. Select Start voice chat in an existing task, or Start new voice chat in a new chat or task.
  3. The first time: allow microphone access, choose a voice, and review screen context on macOS.
  4. Start talking. Select Stop voice chat when you finish.

For faster access, set a shortcut in Settings > Voice > Voice chat hotkey.

One behaviour is still in motion and worth knowing about: voice in existing Codex tasks is rolling out. When available, you can start talking in a task that began with typed messages — voice uses that task's conversation and selected model to carry out your requests, so you can discuss its progress or change direction without starting another task.

If Start voice chat is not available in an existing task, the documented order is to update the desktop app and the Codex host running the task; availability also depends on your account and workspace. You can still start a new voice chat where supported, or use voice dictation to enter prompt text. To resume an earlier voice chat, open it and select Start voice chat.

Interrupting is a supported gesture

ChatGPT Voice supports natural turn-taking. You can interrupt ChatGPT during a response, ask a follow-up, or change direction. If ChatGPT starts work, keep talking to check progress or steer the task.

This is a genuine difference from typing, where you generally wait for the response to finish. In voice, cutting in is designed behaviour, not interference.

Coordinating several tasks

This is the most underrated part of the feature. Voice can:

  • start separate tasks for longer work
  • check existing tasks
  • send follow-up instructions
  • bring progress, blockers and results back to your voice conversation

The three documented examples are the shape of it:

  • "Review today's launch brief and summarize decisions that need approval."
  • "Start a Codex task to run the tests and investigate anything that doesn't pass."
  • "Check active tasks and summarize anything blocking progress."

There is a further move: you can ask to speak to another Codex task, then ask to return to the previous one. Say "Let me talk to the task reviewing the tests", then "Take me back to the previous task". The destination task must support voice and be available on a connected host.

Permissions need no separate thought: ChatGPT Voice follows the same permissions as the tasks it directs. Talking does not route around your approval and sandbox settings.

Screen context: useful, with a cost worth reading

On macOS, turn on Screen context in Settings > Voice, then say "take a look at this". ChatGPT can take an appshot of your frontmost window and use it as context.

The documentation attaches three cautions in the same breath, all worth taking literally:

  • An appshot can include the window's image and accessible text, including content outside the visible scroll area.
  • macOS may request Screen & System Audio Recording and Accessibility permissions.
  • Avoid sharing windows that contain sensitive information, including text outside the visible scroll area.

Read the first and third together: what you can see in a window is not the same as what the appshot carries. A long document scrolled to the middle may hand over the parts you cannot see.

Your organization can disable the capability.

Limits and troubleshooting

Only one voice chat can be active across the ChatGPT desktop app at a time.

Usage has two separate meters, which are easy to conflate:

  • Voice conversations use a separate, plan-dependent allowance measured in rolling five-hour windows.
  • Tasks started through Voice continue to use your Codex usage budget.

ChatGPT notifies you when you reach either limit.

If you cannot start a voice chat, the documented order is: confirm that ChatGPT Voice is available for your plan, rollout and workspace; then check microphone permissions; then check whether a voice chat is already active in another app window. If screen context is not available, check Settings > Voice, Appshots permissions, and your organization's restrictions.

How to do it

  1. Open the Codex task you want to discuss in the ChatGPT desktop app, or start a new chat or task.
  2. Select Start voice chat in an existing task, or Start new voice chat in a new chat or task.
  3. The first time, allow microphone access, choose a voice, and review screen context on macOS.
  4. Start talking, and select Stop voice chat when you finish.
  5. Set a shortcut in Settings > Voice > Voice chat hotkey for faster access.
  6. To let it see something on screen, turn on Screen context in Settings > Voice on macOS, then say "take a look at this".

On Windows

  1. The documentation describes ChatGPT Voice as a ChatGPT desktop app capability on Plus, Pro, Business, Edu and Enterprise plans, without drawing a distinction by operating system.
  2. Screen context, however, is documented specifically as a macOS capability, so that section does not apply on Windows.
  3. If voice is missing from your desktop app, follow the documented order — confirm plan, rollout and workspace availability, check microphone permissions, and check whether a voice chat is already active in another app window.

On mobile

  1. The documentation states you can use ChatGPT Voice through Remote on iOS after pairing your phone with a desktop host.
  2. Mobile use therefore depends on a paired desktop host rather than the phone providing the capability itself.
  3. That passage names iOS only; this tutorial makes no claim about Android behaviour.

Use cases

  • Asking it to review the day's launch brief and summarise the decisions that need approval, while you are away from the keyboard.
  • Starting a Codex task by voice to run the tests and investigate anything that fails.
  • Asking about active tasks and having it summarise whatever is blocking progress.

Common mistakes

  • Confusing ChatGPT Voice with voice dictation. The first is a live conversation with ChatGPT; the second only turns speech into prompt text before sending it.
  • Assuming the feature is broken when Start voice chat is missing from an existing task. Voice in existing Codex tasks is rolling out and needs both the desktop app and the Codex host running the task to be up to date.
  • Using screen context on a window containing sensitive information. An appshot can include text outside the visible scroll area.
  • Trying to run two voice chats at once. Only one can be active at a time.

FAQ

How is ChatGPT Voice different from voice dictation?
The split is documented directly — use ChatGPT Voice for a live conversation with ChatGPT, and voice dictation when you only want to turn speech into prompt text before sending it.
Can it manage several tasks at once?
Yes. ChatGPT Voice can start separate tasks for longer work, check existing tasks and send follow-up instructions, and it brings progress, blockers and results back to your voice conversation so you can keep talking while work continues. You can also ask to speak to another Codex task and then return — "Let me talk to the task reviewing the tests", then "Take me back to the previous task". The destination task must support voice and be available on a connected host.
Is screen context safe to use?
It requires judgement about scope. On macOS, once you turn on Screen context in Settings > Voice, ChatGPT can take an appshot of your frontmost window and use it as context. The documentation notes that an appshot can include the window's image and accessible text, **including content outside the visible scroll area**, that macOS may request Screen & System Audio Recording and Accessibility permissions, and that you should avoid sharing windows containing sensitive information. Your organization can disable the capability.
Are there usage limits?
Two, and they are separate. Voice conversations use a separate, plan-dependent allowance measured in rolling five-hour windows. Tasks started through Voice continue to use your Codex usage budget. ChatGPT notifies you when you reach either limit.
What permissions does it have?
ChatGPT Voice follows the same permissions as the tasks it directs in Chat, Work and Codex in the desktop app. Talking does not bypass your existing approval and sandbox configuration.

Official sources

These are the pages this tutorial is checked against. Follow them if you need the vendor's exact wording.

Source status