Voice dictation in Google Docs: 2026 guide.

Google Docs has built-in voice typing. Open any Doc in Chrome, go to Tools, then Voice typing (or press Cmd+Shift+S on Mac, Ctrl+Shift+S on Windows), and start talking. It's free and fast for short documents. Where it falls short: no formatting, no personal vocabulary, and it works only inside Google Docs. Switch to Gmail or Slack and it's gone.

Last updated: July 2026

How Google Docs Voice typing works

Open a Google Doc in Chrome. Chrome is the most reliable browser for this. Google lists the latest versions of Chrome, Edge, and Safari as supported, but Chrome is still the safest bet. Click Tools in the menu bar, then select Voice typing. A microphone icon appears on the left side of your document. Click it, and when it turns red, start speaking. Your words appear at the cursor in real time.

You need to say your punctuation out loud. "Period" inserts a period. "Comma" inserts a comma. "New paragraph" starts a new paragraph. It takes some getting used to because it breaks conversational flow: instead of just talking, you're thinking about where the periods go while you speak. Voice typing also supports voice commands for basic formatting ("bold," "italics," "undo") and navigation ("go to end of line," "select previous word"), though these work inconsistently depending on your language and browser version.

The feature supports 100+ languages and dialects. It requires an internet connection, since audio is processed on Google's servers. And it's completely free with any Google account.

When built-in Voice typing is enough

If you're writing a short document, drafting notes for yourself, or brainstorming ideas in a Google Doc, the built-in tool does the job. It's free, it's already there, and it doesn't require installing anything. For a student dictating an essay draft, a writer capturing rough thoughts, or someone who occasionally needs to speak instead of type, it's a reasonable starting point.

If this is all you need, save your money. The built-in tool is right for you.

Where built-in Voice typing falls short

The limits show up fast once you try to use Voice typing as part of a daily workflow.

It's Docs-only. Voice typing works inside Google Docs and nowhere else. The moment you switch to a Gmail compose window, a Slack message, a Notion page, or any other app, you're back to typing. For someone who writes across four or five apps every day, a dictation tool that only works in one of them doesn't solve the problem.

No formatting intelligence. Voice typing produces raw text. It doesn't structure your output for the type of document you're writing. You get a stream of words with whatever punctuation you remembered to say out loud, and you format it yourself. There's no distinction between a formal email and a quick note: the output is the same regardless of context.

No personal vocabulary. Voice typing doesn't know your colleagues' names, your project names, your company's acronyms, or any of the terminology specific to your work. "Marcos" becomes "Marcus." "APAC" becomes "a pack." You correct the same words every time you dictate.

You have to speak your punctuation. Saying "period" and "comma" mid-thought interrupts the flow of dictation. It's a learned skill, and some people get comfortable with it, but it's a fundamentally different experience from speaking naturally and letting the tool handle structure.

Browser limits. Chrome is the most reliable browser. Google lists Edge and Safari as supported too, but performance varies. Firefox is not supported. If you use Firefox as your primary browser, Voice typing is not an option.

Rubil in Google Docs

Rubil is a Chrome extension and Mac desktop app that adds voice dictation across every app you write in, including Google Docs.

In Google Docs, Rubil produces clean document prose: properly structured paragraphs, correct punctuation and casing, and formatting that reads like you typed it. You don't say "period" or "comma" because Rubil handles punctuation and structure from your natural speech. Just talk, stop, and the formatted text appears where your cursor is.

The same voice input in Gmail produces a structured email with a greeting and sign-off. In Slack, it produces a concise message with line breaks and no formality. In Notion, it produces document prose with longer sentences. Rubil detects the app you're in and formats the output to match. Voice typing can't do this because it only exists inside Google Docs.

The Glossary is where Rubil goes deeper. You teach it the correct spelling of your colleagues' names and their @handles, acronyms that should expand in some contexts and stay short in others, project names with strict casing, and expansion triggers for templated text blocks. Everything syncs across Chrome and Mac, so your vocabulary follows you everywhere you write.

Rubil formats your words. It never rewrites or generates them. Your voice, your vocabulary, your way of saying things. Rubil handles the structure.

Other alternatives worth knowing

Voice In is a Chrome extension with a large user base that adds voice dictation to web apps including Google Docs. It handles basic transcription well but has no platform-specific formatting and no personal vocabulary system. If you want simple speech-to-text in the browser beyond Google Docs, it's a reasonable free option.

Aqua Voice is a Mac and Windows desktop app with strong technical vocabulary accuracy. It uses a floating-box model (dictate in Aqua's window, then paste into your Doc) rather than inserting text inline. Good for developers and technical writers. For a full comparison, see our Rubil vs Aqua Voice breakdown.

For a broader look at the voice dictation market, including offline options and legacy tools, see our Dragon NaturallySpeaking alternatives guide. And if your day also runs through email or docs, the voice dictation for Gmail and voice dictation for Notion guides cover how Rubil formats for each surface.

Privacy

Voice dictation means sending your voice somewhere for processing.

Google's built-in Voice typing processes audio on Google's servers per their standard policies. Your audio and the resulting text are handled under Google's broader data processing terms.

Rubil's architecture is built around transient processing. Audio goes to Groq for transcription and is discarded immediately. The transcript is sent to OpenAI for formatting and is not stored on Rubil's servers. No voice files, no transcript history, no recordings saved anywhere. Your Glossary lives in your account in our cloud database (Supabase), encrypted in transit and at rest. Every data processor, Groq, OpenAI, and Supabase included, is named on the privacy page.

FAQ