You can dictate text on this computer without buying anything: Windows 11 has voice typing built in (press Win+H), macOS has Dictation (press the microphone key twice, or Fn twice on newer keyboards), and Google Chrome and Microsoft Edge both support voice input through their speech engines. The catch is that each tool behaves differently in where it works, what punctuation it inserts, and whether it needs an internet connection. Knowing which method fits your software saves you from installing a third-party app you don't need.

How Windows 11 Voice Typing Works From the Win+H Shortcut

Microsoft's built-in tool is called Voice Typing in Windows 11, and it replaced the older Windows Speech Recognition panel that still exists in Windows 10. You trigger it with Win+H on any text field, then start talking. The feature appears in Microsoft's Windows 11 accessibility documentation as part of the voice access family of tools.

The tool listens only while the cursor sits in a text box that accepts keyboard input. Click into Notepad, Word, a browser address bar, or a comment field, then press Win+H and a small floating bar appears with a microphone button. Speak, and the words land where you'd expect. Press Win+H again or click the microphone to stop.

Punctuation works through spoken commands rather than automatic insertion. Saying "period," "comma," "question mark," "new line," or "new paragraph" inserts those characters, and Microsoft's support documentation lists the full command set. Saying "delete that" removes the last phrase, which helps when the engine mishears a word.

Windows 11's version adds automatic punctuation as a toggle in the voice typing settings, so you can let the engine guess sentence endings instead of speaking them. Older Windows 10 builds lack that switch entirely, which is the biggest practical difference between the two. If the toolbar never appears, check Settings > Privacy & security > Speech and confirm online speech recognition is on, because the feature depends on it.

Mac Dictation: The Microphone Key, the Fn Key, and What Breaks It

Apple's Dictation lives in System Settings > Keyboard > Dictation on macOS Ventura and later, and in System Preferences > Keyboard > Dictation on older releases. Turn it on, and macOS asks you to confirm the language and whether to use enhanced dictation, which downloads an offline model for larger languages. The default shortcut is pressing the microphone key twice, or pressing Fn twice on keyboards that lack one.

Once enabled, place the cursor in any text field and hit the shortcut. A microphone icon appears near the cursor, and speech becomes text as you talk. Apple's macOS user guide documents the available spoken commands, including "new line," "period," and "cap" to capitalize the next word. Press the shortcut again or click Done to finish.

Dictation can run offline once the enhanced model is downloaded, which matters on planes and in offices with locked-down networks. Apple's documentation notes that some languages and features still require a connection, so offline support isn't universal across locales. Check System Settings > Keyboard > Dictation > Languages to see which model you have installed.

If the shortcut does nothing, three culprits cover most cases. Another app may have claimed the Fn key, the microphone permission for the current app may be denied under Privacy & Security > Microphone, or dictation may be switched off entirely after a macOS update. Re-check the toggle before assuming the hardware failed.

Chrome and Edge Voice Input Without a Desktop App

Google Chrome doesn't ship a system-wide dictation tool, but Google Docs and Google Slides include Voice Typing under Tools > Voice typing, which opens a microphone panel with a language picker. Google's Workspace help documentation describes the feature and its command list, including "period," "comma," and "new paragraph." Chrome then prompts for microphone access the first time.

Microsoft Edge goes further with its own voice typing command, exposed through the same Win+H shortcut on Windows because it relies on the operating system's dictation layer. In Edge on Windows, pressing Win+H over a text field in a webpage works identically to Notepad. That makes Edge the browser of choice for people who dictate into web forms all day on a Windows machine.

Google Docs voice typing drops the language menu into a panel that supports dozens of locales, and you can switch mid-session. It logs words in the document as you speak, and the tool tracks a "voice level" indicator so you can see whether the mic picks up your voice. If the indicator stays flat, the input device is wrong, not the language.

Neither browser tool records audio for later playback. Everything converts to text in real time, so a dropped connection mid-sentence leaves a gap you have to retype. Google's documentation warns that voice typing works best in a quiet room with a headset microphone rather than a laptop's built-in array.

Free Third-Party Apps and How They Compare

Tool Platform Price Offline Live typing into any app Notable limit
Windows Voice Typing Windows 11 Included Needs online speech on for many features Yes Punctuation commands must be spoken or toggled
macOS Dictation macOS Included Yes with enhanced model Yes Enhanced model download required per language
Google Docs voice typing Browser Included with a Google account No No, Docs only Stops when the tab loses focus
Windows Speech Recognition Windows 10/11 Included Yes after training Yes Requires a setup wizard and voice training
Dragon Professional Windows Paid license Yes Yes Costs more than most casual users need

Dragon by Nuance remains the reference product for professional dictation, and its price puts it firmly in the specialist bracket. Nuance's product pages list per-license pricing that changes by edition and region, so confirm the current figure on the maker's site rather than trusting a reseller. For occasional email and notes, the free built-in tools cover the need.

The comparison hides one real trap: Windows Speech Recognition and Voice Typing are different features with different setup paths. Speech Recognition, found via Settings > Accessibility > Speech, requires a training session where you read sample sentences aloud. Voice Typing needs no training but depends on online speech processing for the best accuracy. Users who follow an old tutorial often end up in the wrong panel.

Pick Voice Typing for quick dictation anywhere in Windows. Pick Speech Recognition only if you want a fully offline, trainable engine or you're on a Windows 10 machine that lacks the newer toolbar. Everyone else can skip both and use the browser tools inside Google Workspace.

Getting Punctuation, Capitalization, and Commands Right

Spoken punctuation is the dividing line between a usable transcript and a wall of unpunctuated text. Windows, macOS, and Google Docs all accept the same core vocabulary: "period," "comma," "question mark," "exclamation point," "colon," "semicolon," "open quote," "close quote," and "new paragraph." Each vendor's help center publishes the complete list, and the lists differ slightly at the edges.

Capitalization is handled automatically at sentence starts in all three tools. For mid-sentence capitals, say "cap" followed by the letter on macOS, or use the same command in Windows. Google Docs accepts "capitalize" as a spoken command in its documented list. Names and acronyms usually come through correctly without intervention.

Corrections are where the tools diverge. Windows Voice Typing supports "delete that" and "select that," letting you fix the last phrase by voice. macOS Dictation lets you say "replace" followed by the target word, though the command set is narrower. Google Docs has the fewest correction commands, so plan on fixing errors with the keyboard.

Accuracy depends heavily on microphone quality and background noise. An internal laptop mic in a quiet room typically produces usable text; a shared office or a café pushes error rates up noticeably. A wired or USB headset with a noise-cancelling element is the single biggest upgrade for anyone dictating more than a paragraph a day.

A Worked Example: Dictating a 500-Word Email in Under Nine Minutes

Consider a realistic job: a 500-word client email typed by voice. Speaking at a normal pace of about 150 words per minute produces the raw words in roughly three and a half minutes. Add punctuation commands at sentence ends, and speaking time rises to about four and a half minutes because each command costs a beat.

Correction time is the variable nobody budgets for. At a 95 percent word accuracy rate, a 500-word draft contains around 25 errors to fix, and keyboard correction runs about 10 to 15 seconds per error once you've located it. That adds four to six minutes. Total: roughly nine to eleven minutes for a clean draft.

Typing the same email at 40 words per minute takes about 12.5 minutes before any thinking time. So voice wins, but narrowly, and only if accuracy stays high. At 85 percent accuracy the error count doubles to about 75 fixes, correction time balloons past 12 minutes, and the keyboard becomes faster. Quiet room, good mic, familiar vocabulary: that's the condition under which dictation pays off.

The math also explains why short messages rarely benefit. A two-sentence reply takes 20 seconds to type and 30 seconds to dictate plus corrections. Voice pays off above roughly 150 to 200 words, where the fixed cost of starting the tool and fixing the first errors gets amortized across a longer draft.

Fixing the Microphone When Dictation Types Nothing

The most common failure is a dead microphone in the operating system, not the dictation tool. On Windows, open Settings > System > Sound > Input and watch the level meter while you talk. A flat meter means Windows isn't hearing you, and no dictation feature will work until that's fixed. On macOS, check System Settings > Sound > Input for the same meter.

Permissions come second. Windows 11 has a master microphone switch under Settings > Privacy & security > Microphone, plus per-app toggles below it. macOS has an equivalent list under Privacy & Security > Microphone. If the dictation app isn't listed or its switch is off, it receives silence even though the hardware works.

Third, check the selected input device. Laptops with a plugged-in headset sometimes route audio to the headset's mic while the dictation tool expects the internal array, or the reverse. Set the correct device as the system default rather than per-app, because the dictation layer usually reads the system default.

Fourth, language mismatches produce nonsense rather than silence. If Windows is set to US English and you speak UK English, the engine still transcribes but misspells words like "colour." Check the language in the dictation settings panel, not the system display language, because they're separate controls on both platforms.

Where Dictation Fails: Accents, Names, Codes, and Noisy Rooms

Dictation engines handle conversational English well and struggle with anything outside the training distribution. Product codes, alphanumeric serial numbers, and email addresses are classic failures because the engine hears them as words. Spelling out letters one at a time using the NATO alphabet helps on Windows and macOS, but Google Docs handles it less gracefully.

Strong accents lower accuracy in ways that vary by engine and locale. The same speaker may get better results from Apple's engine than Microsoft's, or the reverse, because the acoustic models were trained on different data. Testing both on a sample paragraph costs two minutes and tells you which to trust for a long document.

Names of people and places are the second weak spot. Adding a name to the system dictionary, where the tool supports custom vocabulary, fixes repeated errors. Windows Speech Recognition has a dedicated dictionary you can populate; Voice Typing and the browser tools mostly don't, which is a real limitation for anyone dictating legal or medical text with unusual proper nouns.

Background noise degrades everything. A fan, an air conditioner, or a nearby conversation raises the error rate even with a decent mic. Noise-cancelling headsets help, but the room matters more. Dictating in a closed office versus an open-plan desk is the difference between a five-minute draft and a fifteen-minute cleanup.

Privacy: What Gets Sent to the Cloud and What Stays Local

Cloud processing is the default for most of these tools, and that has consequences for confidential text. Windows online speech recognition sends audio to Microsoft when enabled, as documented in Microsoft's privacy statements. Turning it off pushes you toward the older Speech Recognition engine, which runs locally but needs training and accepts fewer commands.

macOS Dictation behaves similarly. Apple's support documentation states that when enhanced dictation is enabled and the language model is downloaded, processing can occur on-device. With enhanced dictation off, audio goes to Apple's servers. For anyone dictating client information, the toggle matters more than the feature list.

Google Docs voice typing sends audio to Google's servers every time, with no offline mode. Google's Workspace terms govern the data, and enterprise accounts have different retention rules than free accounts. If your organization blocks cloud transcription, Docs voice typing is out regardless of convenience.

Local-only options exist but cost money. Dragon Professional can run offline, and Windows Speech Recognition is fully local after setup. For regulated industries, matching the tool to the compliance policy comes before matching it to your typing speed. Confirm your organization's rules with whoever owns data policy before dictating anything sensitive.

FAQ

Does Windows voice typing work in every app?

No. It types into any field that accepts keyboard input, but some elevated or protected windows block the Win+H toolbar. Apps running as administrator on Windows 11 often refuse the shortcut until you run the dictation tool at the same privilege level. If Win+H does nothing in a specific program, test it in Notepad first to isolate whether the app or the system is at fault.

Can I dictate offline on a Mac?

Yes, once the enhanced dictation model finishes downloading under System Settings > Keyboard > Dictation. The download is per language and can run to several hundred megabytes. Offline mode supports the core commands, but some newer features and less common languages still require a connection, so test in airplane mode before relying on it during travel.

Why does macOS dictation stop after 60 seconds?

Older macOS releases enforced a time limit on continuous dictation, and some third-party apps still inherit a timeout. The current behavior depends on the macOS version and whether enhanced dictation is active. Speaking a brief pause, then resuming, often resets the session. Check Apple's macOS user guide for your specific version rather than assuming the tool is broken.

Is Google Docs voice typing free?

Yes, it comes with any Google account and requires no paid Workspace plan, though Google's terms differ between consumer and business accounts. The tool only works inside a Google Docs or Slides tab in Chrome, and it stops when you switch tabs. For dictation into email, forms, or desktop software, you need one of the operating-system tools instead.

Which is more accurate, Windows or Mac dictation?

Neither wins universally, because accuracy depends on your accent, microphone, and the vocabulary you use. Running the same 100-word paragraph through both takes two minutes and gives you a personal answer that beats any general benchmark. Microsoft and Apple both update their speech models with OS releases, so a verdict from two years ago may not hold today.

How much does professional dictation software cost?

Paid tools like Dragon Professional sit well above the free built-in options, with per-license pricing that varies by edition and region. Nuance publishes current figures on its own site, and resellers often differ. Buy it only if you dictate several thousand words a day or need offline processing for compliance, because the free tools cover casual use.

Can I dictate into a website form?

Yes, if the form's text fields accept keyboard input and your browser allows the dictation layer. On Windows, Win+H works over Edge and Chrome fields. Google Docs voice typing does not, because it's scoped to the Docs editor. If a form field rejects text, the site's script is likely blocking programmatic input, not the microphone.