Voice to text in Google Docs

Open Tools, click Voice typing, start talking

No install, no extension, no account upgrade. Voice typing ships inside Google Docs already.

The whole thing takes about thirty seconds the first time, and ten on every visit after. If the microphone box never appears, you are almost certainly in Firefox or an older browser build. Jump to the troubleshooting section below.

One detail worth setting before you start: the language drop-down above the microphone. Google Docs supports 100+ languages and dialects there, including a long list of regional variants of English, Spanish, and Arabic. Pick the right one and accuracy jumps the moment you do. Leave it on a language you do not speak and you will get word salad that reads like a bad subtitle track.

The punctuation commands and the shortcut you need

Here is the part that surprises people. Google Docs does not add punctuation for you. You say it.

To end a sentence, you say "period." For a pause, "comma." You also have "exclamation point," "question mark," "new line," and "new paragraph." So dictating two sentences sounds like this out loud: "the meeting is confirmed period send the agenda by Friday period." It feels strange for about four minutes, then it becomes muscle memory, the way typing a semicolon once did.

A wider set of editing commands also exists: "select paragraph," "italics," and similar formatting and navigation phrases. Those work in English only, and only when both your account language and your document language are set to English. Punctuation commands are more forgiving and work across more languages; the formatting commands are the English-locked ones.

A keyboard shortcut floats around too. Many users and third-party guides report it as Ctrl+Shift+S on Windows and Cmd+Shift+S on Mac to toggle voice typing without reaching for the menu. Worth a mention because Google's current help page does not list a shortcut at all, so treat it as "widely reported and usually works" rather than gospel. If it does nothing on your machine, the menu route always works.

The manual-punctuation thing is the cleanest line between Google's tool and the newer AI dictation tools. Saying "comma" out loud is the price of a free feature that has not changed much in years. It is not wrong. It is just 2015 sitting quietly inside a 2026 product.

It runs in a browser tab, not on your computer

The next section decides whether the built-in tool is enough for you, so it gets a picture of what you are looking at.

Voice typing is a feature of Google Docs, which means it is a feature of the browser tab Google Docs is open in. There is no desktop app. Close the tab and the microphone goes with it. The earlier "Chrome only" reputation is out of date — Google's help page now lists Chrome, Edge, and Safari. Firefox is still left out.

What that browser-lock costs you shows up the moment your writing leaves the document. Say you draft a project update in Docs by voice, then you need to reply to a comment in your email, paste a note into Slack, and fill a cell in a spreadsheet. The microphone cannot follow your cursor into any of those. It only hears you inside the Docs tab. For a lot of people that is fine — they write in Docs and nowhere else. For the rest, it is a wall you hit several times a day.

Then there is the mobile question. On a phone, the desktop Voice typing menu does not appear; the Docs mobile app leans on your phone keyboard's built-in dictation microphone instead. That keyboard mic is decent, but it is your operating system doing the work, not Google Docs.

When Google Docs voice typing stops hearing you

Voice typing breaks in a small number of predictable ways. Ranked by how often I have watched them happen.

• It is missing from the Tools menu. You are in an unsupported browser. Firefox is the usual culprit; an ancient Chrome build is the runner-up. Open the same document in current Chrome, Edge, or Safari and the menu item reappears.
• The microphone box is there but nothing is typed. The browser does not have microphone permission, or the wrong input device is selected. Check the little camera/microphone icon in your browser's address bar and confirm the document tab is allowed. Then check your operating system's sound settings to make sure the right microphone is the default — the built-in laptop mic and a plugged-in headset fight over this.
• Words run together or come out garbled. This is almost always audio, not software. A cheap USB microphone does more for accuracy than any setting you can change, usually more than a model upgrade does. Background noise, a fan, an open window: all of it lands in the transcript.
• It stops on its own after a minute or two. Voice typing pauses when it stops detecting speech for a stretch, or when the tab loses focus. Click the microphone again to resume. If you switch to another window to read something while you dictate, that switch is what is muting you.
• Words appear in the wrong place, or commands get typed as text. If you say "new paragraph" and the words "new paragraph" land in your document instead of a line break, your account or document language is not set to English, and the editing commands are English-only. Punctuation like "comma" and "period" is more forgiving across languages; the formatting and navigation commands are the strict ones. Check both language settings if commands keep showing up as literal text.

If the document tab works but a different app does not hear you at all, that is not a Google Docs problem. That is the browser-lock again. No troubleshooting step fixes it, because it is working as designed.

If the microphone box never starts listening at all, see our full guide to why Google Docs voice typing stops working.

Dictating into Docs without living inside a browser tab

A system-wide tool changes the shape of the problem. Whisper by Remskill is a desktop app, not a browser extension and not a Docs add-on. You press a hotkey anywhere, speak, release, and the text is pasted at your cursor — in the Google Docs tab, in your email, in Slack, in a spreadsheet cell, in any text field on the machine.

The default hotkey is Ctrl+Space on Windows. On Mac it is a push-to-talk: hold Command+Option together to record, release either key to stop. Both are changeable in settings if they clash with something you already use. The flow is the same on either platform — hold, talk, release, the words land where your cursor is. You can see the full walkthrough in our guide to how Whisper works, and the platform-specific setups for voice to text on Mac and voice to text on Windows.

A few things the browser tool structurally cannot do, that a desktop app can:

• Punctuation is automatic. No saying "comma" out loud. The transcription adds it for you, so dictation sounds like talking, not like reading stage directions.
• It works offline. The local pipeline runs fully on your machine through a pure-Rust engine, no audio leaves the laptop. Your half-drafted resignation letter does not take a detour through anyone's server.
• It follows your cursor everywhere. The same hotkey that fills a Docs paragraph fills a Gmail reply or a Notion page thirty seconds later.

A word on models, because people ask. Locally you pick the engine yourself — the choice is yours, not a default we picked for you. NVIDIA's Parakeet is the fastest local option: about 600 MB, 5 to 10 times faster than Whisper on a CPU, covering English plus 24 European languages, 25 in total. Pick it if you mostly speak English and you want speed.

The local Whisper models go wider. The multilingual builds cover 99 languages including auto-detect, plus translate-to-English. Pick those if you need other languages or translation. The English-only Whisper builds are locked to English by design, for the best accuracy when English is all you write. And if you want the latest OpenAI quality with web access, Cloud mode runs on your own OpenAI key, covering the roughly 57 languages OpenAI officially lists. Three paths, you choose. The app does not pick a default for you, which I learned the hard way is exactly what people want — the first build shipped with a model already chosen, and the first thing users did was go change it.

That last point is the lunchbox test for me. A Tuesday evening, making sandwiches and the same yogurt the younger one will refuse to eat, and the school wants the trip slip answered by eight. I grab the laptop one-handed, hold the hotkey, and dictate the reply between cucumber slices — stopping once to ask how to spell the teacher's name, once to answer why the moon is sometimes not there. The email goes out, the lunchboxes get made. The point is not that this is Docs. The point is that my cursor was in an email client, then a calendar, then a notes app, and the same key worked in all three. Voice typing would have heard the first one and gone silent for the rest.

When to skip our app and use Google's built-in tool

I will say this plainly because the framework I write by demands it, and because it is true. If you write inside Google Docs and nowhere else, and you are sending short bursts — a paragraph, a comment, a quick note — use Google's Voice typing and do not install anything. It is free, it is already there, it supports 100+ languages, and saying "period" a few times costs you nothing but dignity.

The line where a system-wide tool starts to matter is roughly the 200-word mark, or the moment your dictation needs to leave the document. If you are drafting long-form, switching between three apps an hour, working offline on a train, or handling text you would rather not send to a server, the browser tab becomes the bottleneck. Below that line, Google has you covered. Be honest with yourself about which side you are on.

If you have already crossed that line, here is a Google voice typing alternative that dictates anywhere — including offline.

One more built-in option people forget. Both Windows and macOS have their own system-level dictation. On Windows it is the Win+H voice typing panel; on Mac it is the operating-system dictation you can bind to a key. Those at least work outside the browser, unlike Google's. They are not as accurate or as controllable as a dedicated tool, but they are free and already installed — worth trying before you reach for anything paid.

Google Docs voice typing is free, built in, and fine for a paragraph at a time inside a browser tab. The moment your dictation needs to leave that tab — into email, into a chat, into a spreadsheet, offline on a train — the microphone stays behind, because it was never on your computer to begin with. That is not a flaw to fix. It is the edge of what a browser feature can reach. My seven-year-old once dictated a whole email to her grandmother and then asked why it would not work in her drawing app. Reasonable question. The answer is a hotkey that does not care which window you are in.