Voice to text on a Mac: four ways, compared

Your Mac can already type what you say. Speech to text on a Mac comes in four forms: two built into the system, one that costs money, and one that is free but lives in the terminal. They are not interchangeable: one is meant for a paragraph at a time, one is meant for running the whole computer by voice, and the other two exist because the first two leave gaps.

The short answer

Dictation is built into macOS. It costs nothing, it is already installed, and it handles short bursts of text well.

Voice Control is also built in, but it is an accessibility feature for operating the Mac by voice. It includes dictation, and it switches regular Dictation off while it runs.

A paid dictation app buys you a hotkey that works in any app, a place to add the words you actually use, and, depending on the app, recognition that stays on your machine.

Whisper run by hand is free and accurate. You feed it recordings and get text back. It will not type into Slack for you.

Built-in Dictation

Apple’s own speech to text has shipped with macOS for years, and most people never turn it on.

To turn it on, open the Apple menu, then System Settings, then click Keyboard in the sidebar. Go to Dictation and switch it on.

To start dictating, press the microphone key if your keyboard has one, use the shortcut you set in that same pane, or choose Edit, then Start Dictation, in whatever app you are typing in.

What it handles well

There is no limit on how long you can talk. Apple’s documentation says dictation stops automatically when no speech is detected for 30 seconds, which is a silence timeout rather than a cap on length. Older versions of macOS did cap dictation at around a minute, which is where a lot of outdated advice comes from.

On Macs with Apple silicon, voice inputs and transcripts for general text dictation are processed on the device rather than sent to Siri servers. Apple tells you to check Keyboard settings to confirm this, which is worth doing once if it matters to your work.

Voice typing on a Mac handles punctuation two ways. You can say the name of the mark, such as “exclamation mark”, and in supported languages macOS inserts commas, periods and question marks on its own. On Apple silicon you can also keep typing while dictation is running, without stopping it first.

Where people get stuck

There is no vocabulary list. If you dictate the name of a colleague, a drug, a library or a client, you get whatever the model heard, and you correct it by hand every time. For most people this is the single reason they go looking for something else.

Language support is a list rather than a guarantee. Apple publishes which languages you can dictate in on its feature availability page for the current version of macOS, and if yours is missing, none of the above helps you.

Formatting is basic. You can say “new line” and “new paragraph”, but there are no rules you can define, so a variable name arrives as three separate words and a phone number arrives as prose.

Voice Control is a different feature

People find Voice Control while looking for better dictation and come away confused, because it was built for something else. Apple describes it as a way to navigate the desktop and apps, interact with what is on screen, dictate and edit text, and more.

It lives elsewhere in settings: Apple menu, then System Settings, then Accessibility in the sidebar, then Voice Control. You may have to scroll to find it.

One sentence in Apple’s documentation matters more than the rest: “When Voice Control is on, you use Voice Control to dictate text; standard macOS Dictation isn’t available.” The two do not stack, so you are choosing between them rather than combining them.

Voice Control has three ways of getting text in. Dictation mode is the default, and anything you say that is not a command becomes text. Spelling mode goes character by character and accepts phonetic codes, so “Alfa Bravo Charlie” produces abc. Command mode ignores everything that is not a command.

It edits, too. Say “Replace cat with dog” and it does exactly that. If your goal is to use a Mac without a keyboard, whether because of injury, disability, or a wrist in a cast for six weeks, this is the right tool, and the dictation inside it is a bonus. If your goal is to write faster while still using a keyboard and trackpad, it is more machinery than the job needs.

Paid dictation apps

This category exists because of the gaps above. We compared eight of them, with prices and where each one recognises speech, in best dictation apps for Mac. What you are buying, roughly in the order people care about it: a global hotkey that drops text into whatever field your cursor is already in, so you stop switching windows to dictate; a dictionary where you add your own names and jargon; post-processing that turns speech into something closer to what you would have typed; file transcription rather than live speech only; and a choice of recognition model, which matters if you speak a language where one model is clearly better than another.

Before paying for any of them, ask where the audio goes. Some of these apps capture on your machine and send the recording elsewhere for recognition, which is a normal engineering decision and usually documented somewhere. It is also not what people assume when they read “runs on your Mac”. Capturing locally and recognising locally are two different claims, and only the second one means the recording never leaves. If an app talks about privacy without saying where recognition happens, treat that as the answer. Ours spells out the route the audio takes, and you should expect the same from anyone else.

Disclosure: we make one of these. Kekoso costs $29 once instead of monthly, and everything it does happens on your Mac: dictation into any app, audio and video files, YouTube links, meeting recordings with the microphone and system audio on separate tracks, a vocabulary you fill in yourself, and a local MCP server your coding agent can call. Recognition runs on the Neural Engine, the audio is discarded once the text lands, and the licence covers 3 machines with a 7-day trial. It needs macOS 14 or later on Apple silicon, which rules out every Intel Mac and is the main reason to pick something else.

Whisper, run by hand

Whisper is OpenAI’s open speech recognition model. You can run it on your own machine for free, and there are builds tuned for Apple silicon that are fast enough to be practical.

You get accuracy, no subscription, no account, and nothing leaving your computer.

You do not get any connection to the way you work. There is no hotkey, and it does not type into other apps. You record a file, run a command, get a transcript, and paste it where you needed it. For a folder of interviews, that is a perfectly good workflow. For answering an email, it is absurd.

Setup is an evening, and it stays a small ongoing chore: model files to manage, command flags to remember, an occasional broken dependency after a system update. Some people enjoy that. If you are not one of them, that preference is most of what a paid app is selling you.

Side by side

	Dictation	Voice Control	Paid app	Whisper by hand
Price	Free	Free	One-time or subscription, varies	Free
Types into any app	Yes	Yes	Yes	No
Recognition on device	On Apple silicon, per Apple	Not documented	Varies by app	Yes
Your own vocabulary	No	Commands only	Usually	Limited
Controls the computer	No	Yes	No	No
Transcribes files	No	No	Usually	Yes
Setup	Minutes	Minutes	Minutes	An evening, plus upkeep

How to choose

If you dictate a few times a week, use the built-in Dictation. It is already there and it is good enough for a paragraph. Paying for something you touch twice a week makes no sense.

If you write for hours a day, a paid app earns its price back in corrections you stop making, especially if your work has its own vocabulary. Spend a week with the free options first so you know precisely what you are buying.

If you want to run the Mac by voice, use Voice Control, and accept that it replaces Dictation while it is on.

If you have a pile of recordings, use Whisper or an app that handles files. Live dictation tools are the wrong shape for that job, and the options are compared in how to transcribe audio and video to text on a Mac.

If your work is confidential, the only question that matters is where recognition happens, and you should verify the answer rather than trust a landing page. Apple documents on-device processing for general dictation on Apple silicon. For anything else, look for a plain statement about where the audio is recognised, and treat the absence of one as a no.

Questions people ask

Does Mac dictation work without an internet connection?

On Macs with Apple silicon, Apple documents that voice inputs and transcripts for general text dictation are processed on the device rather than sent to Siri servers, and tells you to check Keyboard settings to confirm it for your machine. Other features may still use the network, so check that settings pane rather than assuming.

Is there a time limit on Mac dictation?

There is no cap on how long you can talk. Dictation stops on its own when it detects no speech for 30 seconds.

What is the keyboard shortcut for voice to text on a Mac?

Press the microphone key if your keyboard has one, or use the shortcut you set under System Settings, Keyboard, Dictation. You can also choose Edit, then Start Dictation, inside the app you are typing in.

Can I teach Mac dictation my own words?

Built-in Dictation has no vocabulary list you can edit. Names, acronyms and technical terms are the most common reason people move to a third-party dictation app.

What is the best speech to text app for a Mac?

It depends on how often you dictate. For a few paragraphs a week, the built-in Dictation is already installed and costs nothing. For daily writing with names and technical terms in it, a paid app with a custom vocabulary saves more time than it costs. For a folder of recordings rather than live speech, use a tool built for file transcription.

Is Voice Control the same as Dictation?

No. Voice Control is an accessibility feature for operating the Mac by voice, and Apple's documentation states that while Voice Control is on, standard macOS Dictation is unavailable.