Voice typing in Gmail
Press your hotkey, say the reply, and punctuated text lands in the compose window. No extension to install, no permission on your mailbox, and the same hotkey works in the tab next door.
Does Gmail have voice typing built in?
No. Google’s voice typing is a Google Docs feature and has never worked in the compose window. What you may already have is your operating system’s own dictation — Win+H on Windows, the dictation key on a Mac — which does type into Gmail.
The case for this over either is the punctuation and the reach: full stops and paragraphs are applied from the way you speak rather than dictated as commands, and the same hotkey works in the document, the chat window and the form as well.
The reply, not the draft
Mail is short, conversational and formulaic — the writing speaking gains most on. A fifty-word reply takes about 20 seconds to say and about 75 seconds to type, at the 150 and 40 words a minute those figures assume.
Nothing on your mailbox
No extension, no OAuth grant, no mailbox permission to review. Gmail never learns this exists; the compose window receives keystrokes the way it does from a keyboard.
Subject line and all
Anywhere a cursor sits: the subject, the body, a chip in the recipient field, the reply box under a thread. There is nothing per field to enable.
How it works with Gmail
Three steps, none of them inside Gmail.
Install the desktop client
A native binary for Windows, macOS or Linux, linked to your account once with a device code. No extension, and nothing is granted access to your mailbox.
Open a reply and click into it
The body, the subject, a chip in the recipient field, the inline reply box under a thread. The client types into whatever is focused.
Press your hotkey and speak
Punctuated text arrives in the compose window as you go. Read it when you are done, press send yourself — nothing is sent for you.
Where the time actually goes in a mailbox
Speaking a reply is faster than typing one, and the figures for that are on this page. What they do not cover is the other half of answering mail: reading the thread, working out what to say, and deciding whether to send it at all. Dictation does not touch that half, and a page claiming it halves your mail would be counting the wrong thing.
What it changes is the backlog of replies that are quick to think about and slow to type — the twelve short answers that are still sitting there at six o’clock. Those are the ones where the typing genuinely is the work.
What it does not do
- No mail commands. It cannot send, archive, add a recipient or attach anything — it types where the cursor is and stops there.
- It does not read your mail, and could not: nothing connects to your account, so it has no idea what thread you are in or what was said above.
- No offline mode. Audio is transcribed on our servers, so a train with no signal is a train where this does not work.
- Not with the desktop client. On Android our own keyboard does it, while it is in closed testing; otherwise the WhatsApp bot returns text you can paste, which is a different way of working. On iOS there is nothing from us.
Common questions
Is this a Gmail extension?+
No, and deliberately not. There is nothing to install into the browser, nothing to grant access to your mailbox and nothing installed for an administrator to approve. The client runs on your machine and sends keystrokes to whichever window has focus. What does leave it is the audio, streamed to our servers to be transcribed — which is the part an administrator would actually be reviewing.
How is this different from Windows dictation?+
Win+H types into Gmail too, so the difference is not that one works and the other does not. It is that punctuation and sentence breaks are applied here from the way you speak rather than spoken as commands, and that the same hotkey is the same hotkey on macOS and Linux — one thing to learn rather than one per operating system.
Does it work in the Gmail mobile app?+
No. The client is desktop software, so the answer on a phone is our own Android keyboard while it is in closed testing, the phone’s built-in dictation, or the WhatsApp bot if you would rather send a voice note and paste the text back.
Does it work in Outlook or Thunderbird?+
Yes — and in any other mail client, because none of them is treated specially. The exceptions across the board are programs that deliberately reject synthetic input — some games, some remote-desktop and virtual-desktop sessions, and some hardened enterprise clients.
Can I dictate in another language?+
Yes. The language is detected from what you say rather than set per application, so switching between two languages during a working day needs no setting changed.
The same client types into Slack, Word and Google Docs. Comparing dictation tools first? VibeVoice vs Wispr Flow is the closest match.
Last updated . Competitor figures checked August 2026: prices, platforms, and each vendor’s own description of when its text is finished, which is their account rather than our measurement. All of it changes, so check theirs before buying.