How to dictate emails in Gmail by voice

Imagine a company that builds a beautiful recording booth in one room of its office. Soundproofed walls, a studio-grade microphone, the works. Now imagine that same company forgot to put a microphone in the room where everyone actually talks all day. That is Google with dictation. It built Voice typing into Google Docs, where you write the occasional long document, and it put nothing into Gmail, where you probably type far more words every single day. Same company, same account, and yet the Gmail compose box just sits there in silence.

So can you dictate an email in Gmail? Yes, you can, and this guide shows you exactly how. But the honest headline is that every method you will read about here is something living outside Gmail, typing into the compose box on your behalf. Gmail itself has no ears. Let us walk through the surprising gap, the workaround that works today, its real limits, and the smoother fix if you write a lot of email by voice.

The surprising gap: Docs can hear you, Gmail cannot

Open a document in Google Docs, click Tools > Voice typing, and a little microphone box appears. Click it and start talking, and your words land on the page. It works in Chrome, Edge and Safari (not Firefox), and it is a genuine, built-in Google feature. Google really did build dictation into Docs.

Now open Gmail, hit Compose, and go looking for that same microphone. It is not there. There is no native mic button, no hidden dictation option buried in a menu, nothing. Gmail on the web has no built-in voice typing at all. To dictate an email you have to reach for something that is not part of Gmail: your operating system's dictation, a browser extension, or a third-party tool. This is the whole point, and it is worth saying plainly: Gmail has no dictation of its own.

The gap is strange precisely because Gmail is where the writing volume is. Most people fire off dozens of emails a day and open Docs a handful of times a week. Google put the microphone in the quiet room and left the busy room silent. (A small honesty note: Chrome has recently started adding its own 'Start dictation' right-click option in test builds, but as of mid-2026 it does not actually produce text yet. It does not change the conclusion. Gmail still has no dictation of its own.)

So how do you dictate an email today?

The workaround that works right now is your operating system's own dictation. Both Windows and Mac ship with a system-wide dictation that can type into almost any text field, including the Gmail compose box in your browser. Here is the honest, step-by-step.

1. 1

   On Windows: press Win + H

   Open Gmail in your browser and click Compose. Put your cursor inside the compose box, then press the Windows key + H. A small voice-typing bar appears at the top of the screen. Start talking and your words go into the email. This is Windows' own Voice Typing working inside the browser field, not a Gmail feature. See Microsoft's guide at support.microsoft.com.

2. 2

   On Mac: turn on and trigger Apple Dictation

   Enable it once under System Settings > Keyboard > Dictation. Then click into the Gmail compose box and trigger dictation with your shortcut (by default, press the microphone or Fn key). Speak, and macOS types it into the email. Again, this is macOS' built-in dictation working inside the browser, not something Gmail provides. See Apple's Dictation guide.

3. 3

   On your phone: use the keyboard mic

   In the Gmail app, tap into the message field so the keyboard comes up, then tap the microphone on the keyboard itself (the iOS keyboard or Gboard). Speak and it types. That mic belongs to the phone's operating-system keyboard, not to Gmail. The Gmail app has no dictation of its own either.

One honest caveat on Windows: Win + H works in most browser text fields, but not every web app, and there are Chrome-specific quirks, so it can be a little inconsistent inside Gmail. When it works it is genuinely useful and completely free with software you already own.

The limits of the OS-dictation workaround

The system dictation is a real, usable option. But once you write more than a line or two, you start bumping into the same rough edges, and they are worth knowing before you rely on it for real email.

• You have to speak the punctuation. You literally say 'comma', 'period', 'new paragraph' out loud. Miss one and your email becomes a run-on wall of text you have to go back and fix.
• One language at a time. OS dictation follows a single selected language. Drop a second language, or an English brand name into a Spanish email, and it forces those sounds through the wrong language model.
• It mangles names. Proper nouns, client surnames, product names and brands come out misheard, exactly the words in an email you least want wrong.
• It can cut out when focus shifts. Browser-based dictation can stop when the tab or window loses focus, so clicking away to check something can quietly end your session mid-sentence.

None of these are dealbreakers for a quick reply. But stacked together on a long, careful email, they turn dictation into a narrate-then-repair chore: you say the punctuation, you fix the names, you re-read for the run-ons. The time you saved talking gets eaten by the cleanup.

The smoother fix: system-level dictation that arrives clean

If you write a lot of email by voice, what you actually want is dictation that lands already clean in the compose box: punctuation added for you, names spelled right, and no menu to babysit. That means a system-level tool that sits above the browser and types into any app, with an AI cleanup pass between your voice and the text.

Golem is one tool built exactly this way. It is OS-level, so it works in Gmail on the web and on desktop, and in any other app where your cursor is, editor, chat, terminal, ChatGPT. It runs OpenAI's Whisper (large-v3-turbo) for transcription and then adds an AI cleanup pass, so the text arrives with punctuation already in place instead of you dictating every comma. It supports custom vocabulary, so you can teach it the names and brands you email most and stop seeing them misheard. And because Whisper is a many-language engine, it handles a Spanish email with an English term in it far more gracefully than a single-language OS setting. It needs a login and is free to start with no card. To be clear, Golem is not a Gmail feature or a Google partnership. It is an outside, OS-level tool typing into the compose box, just like Win + H, only smoother.

Golem is not your only option, and it would be dishonest to pretend it is. The free OS dictation above (Win + H, Apple Dictation) is perfectly fine for short replies. Wispr Flow is a polished paid option at around $15 a month. Superwhisper is a strong Mac-first choice with on-device modes if privacy or offline use matters to you. The honest way to choose is by how much email you actually dictate.

So which suits you?

• Fire off a quick reply now and then? Win + H on Windows or Apple Dictation on Mac is enough. Free, already installed, no extra tool.
• Write long, careful emails all day? You want dictation that arrives punctuated and spells names right, so a system-level tool like Golem saves you the cleanup pass.
• Email in more than one language, or with lots of brand and client names? A many-language engine with custom vocabulary handles this far better than a single-language OS setting.
• Care about privacy or working offline? Look at an on-device option such as Superwhisper.
• Want to go further? Read the companion guide on how to dictate in Word and Google Docs, and, since speaking every comma is the main annoyance above, dictation that punctuates itself.