How to talk to text on a Mac

What "talk to text" actually means

On a Mac, talk to text is called Dictation. It's a feature that listens through your microphone, turns your speech into written words, and types them wherever your cursor is, in Mail, Messages, a Google Doc, a text box on a website, anywhere you could normally type. You don't need to open a special app. Once it's on, it works the same way in almost every program on your Mac.

It is not the same thing as Siri. Siri answers questions and controls your Mac. Dictation just converts what you say into text, the way a very fast, very literal typist would.

How to turn on dictation on a Mac

1. Open System Settings. Click the Apple logo in the top-left corner of your screen and choose System Settings. (On an older Mac still running macOS Monterey or earlier, this is called System Preferences instead.)
2. Click Keyboard in the sidebar. Scroll down if you don't see it right away.
3. Turn on the switch next to Dictation. A window pops up explaining that using Dictation may send your voice to Apple to process it. Click Enable to confirm.
4. Wait for the download. The first time you turn it on, your Mac downloads a language file so it can recognize speech. This usually takes under a minute on a normal connection, longer on a slow one.

That's the whole setup. You won't need to repeat it unless you erase your Mac or switch to a different language.

How to use it once it's turned on

• Click into any text field, an email, a document, a chat, a search box.
• Press the shortcut to start listening. On a MacBook from 2021 or later, this is a dedicated key in the top row with a small microphone icon on it, usually sitting where F5 used to be. On an older Mac, or a Magic Keyboard, press the Fn key twice in a row.
• A small microphone icon appears near your cursor. Start talking at a normal pace, the same way you'd talk on a phone call.
• Press the same key (or the Fn key twice again) to stop. On most Macs, dictation also stops on its own after a few seconds of silence.

Words that work as punctuation and commands

Dictation understands a set of spoken words as instructions rather than text to type. Auto-punctuation adds some of these for you automatically, based on how you're speaking, but you can also say them out loud for more control:

• "period," "comma," "question mark," and "exclamation point" insert that punctuation mark.
• "new line" starts a fresh line; "new paragraph" adds a blank line first.
• "cap" capitalizes the next word; "all caps" capitalizes the next word entirely; "no caps" keeps it lowercase.
• "open parenthesis" and "close parenthesis" add those marks around whatever comes next.
• Saying the name of an emoji, like "smiley face emoji" or "thumbs up emoji," inserts that emoji.

These commands feel awkward at first, since it means narrating your own punctuation mid-sentence. Most people either lean on auto-punctuation and skip most of these, or stop using them altogether and clean up punctuation by hand afterward.

Where people get frustrated with it

Apple's dictation is genuinely useful for short bursts of text, and it falls short in a few specific, predictable ways:

• It types everything, including the mess. "Um, so I think, uh, we should probably move the meeting" comes out exactly like that. Dictation doesn't judge which words matter; it transcribes what it hears.
• Punctuation is your job. Auto-punctuation guesses reasonably on short, clear sentences, but longer or more rambling speech often ends up as one run-on block, or with commas in odd places.
• Names and jargon get mangled. A coworker's uncommon name, a product name, an acronym specific to your industry, these often come out wrong, and there's no way to teach Dictation your own vocabulary.
• It doesn't know what you're writing. Dictating an email doesn't produce a greeting, paragraph breaks, or a sign-off; it produces a wall of text you still have to shape into an email.
• It can cut you off. If you pause too long to think, or take a breath at the wrong moment, dictation may stop listening, and you'll notice only after you've kept talking to an empty text field.

Tips that make it work better

A few habits close most of the gap:

1. Talk in full sentences, at a steady pace. Fast, choppy speech with lots of restarts is exactly what produces the messiest output.
2. Say punctuation as you go for anything longer than a sentence or two, rather than relying on auto-punctuation to guess correctly.
3. Dictate in a quiet room. Background noise, a TV, an open window near traffic, a nearby conversation, all of it gets picked up and can turn into stray words in your text.
4. Read it back before you send anything. Treat dictated text the way you'd treat a first draft, not a finished message.

If you'd rather not clean it up by hand

Everything above is free and already on your Mac. The tradeoff is that you're the editor: every filler word, every missing comma, every wrong name is yours to fix after you dictate.

Amlin is a Mac app that adds a cleanup step on top of the same idea. You hold a key and talk, the same as built-in dictation, but instead of typing your words verbatim, Amlin removes the ums and false starts, adds punctuation and capitalization, and shapes the result for where you're typing: an email comes out with a greeting and paragraph breaks, asking for a bullet list produces real bullets, saying "new line" still works, and a requested emoji still lands where you asked for it. It also learns your own names and terms over time, through a custom dictionary shared with Amlin's meeting notes, so the same word that trips up built-in dictation gets spelled right the next time.

The cleanup runs entirely on your Mac, the same as Apple's own on-device dictation. Nothing is sent anywhere, and it's included free on every plan, with no separate limit on how much you dictate. Amlin requires macOS 13 or later on Apple silicon.