There are three ways to dictate on a Mac in 2026: the built-in macOS dictation (free, types your words literally), push-to-talk AI dictation (hold a key, speak, release — polished text appears in the focused field), and fully hands-free voice typing with a wake word or voice-activity detection. Built-in dictation is fine for short literal notes; AI dictation apps like Clavio additionally rewrite your speech into clean, ready-to-send text in the tone you choose, work in every app, and can be driven entirely by voice.
Option 1 — built-in macOS dictation
Press the shortcut you set in System Settings → Keyboard → Dictation (F5 by default on recent Macs) and speak. macOS types your words as-is, punctuation included only when you say it out loud. It is free and private, but it does not clean up filler words, fix grammar, restructure rambling sentences, or adapt tone — what you say is exactly what lands in the text field. For quick literal notes it is enough; for messages, email, and documents most people spend longer editing than they saved by speaking.
Option 2 — push-to-talk AI dictation
An AI dictation app sits in the menu bar and binds a global key. Hold it — for example Globe (Fn) or ⌥ Space in Clavio — talk naturally with all the 'uhm, so, basically' rough edges, release, and the app transcribes your speech, rewrites it into polished text in your chosen tone, and pastes the result into whatever field is focused. That last part matters: it works in every app — mail, chats, browsers, editors, terminals — because the text is delivered at the cursor, not into a separate window you copy from. You speak in any language you use and it survives mid-sentence language switches; you pick a default tone and override it per app, so your chats sound relaxed while your email stays buttoned-up; and a personal dictionary plus snippets keep names, brands, and boilerplate exactly the way you write them.
Option 3 — hands-free: wake word and voice activity
When your hands are busy, two activation styles replace the key entirely. With a wake word, the mic waits for a trigger phrase and records the thought that follows — a verbal gate before anything is captured. With voice-activity detection, dictation starts the moment you begin speaking and stops on the pause. In Clavio both modes are backed by a local voiceprint, so the Mac responds to your voice rather than the room's, and wake-word recognition runs on-device.
Which one should you use?
Use built-in dictation if you rarely dictate and only need literal transcription. Use push-to-talk AI dictation as the daily default — it is the fastest way to turn rough speech into a sendable message. Add hands-free modes when you dictate while cooking, driving a desk full of windows, or simply think better away from the keyboard.
Try the AI route
If you want to try AI dictation, Clavio for Mac is free to download — 3,000 polished words a month, no card, macOS 14+ on Apple Silicon. The onboarding sets up push-to-talk, wake word, and hands-free modes with your own voice.
Keep reading
Common questions
How do I turn on dictation on a Mac?
For built-in dictation, open System Settings → Keyboard → Dictation, turn it on, and press the shortcut (F5 or the 🌐/Fn key by default) to start speaking. For AI dictation, install an app like Clavio and it binds its own key or hands-free trigger.
How do I dictate into any app on a Mac?
An AI dictation app that pastes at the cursor works in every app — mail, chats, browsers, editors, terminals — because it delivers the finished text into whatever field is focused rather than a separate window. Clavio does this system-wide.
Can I dictate on a Mac without holding a key?
Yes. Hands-free modes replace the key: a wake word starts recording when you say a trigger phrase, and voice-activity detection starts when you speak and stops when you pause. Both are keyed to an on-device voiceprint so the Mac responds to your voice.
What's the best way to dictate long text on a Mac?
Use an AI dictation app with its own capture length — built-in dictation cuts off after about 30 seconds. Clavio records a full thought in one pass and then cleans it up into finished text.