Push-to-talk dictation is a model where the microphone is only active while a key is physically held down: press, speak, release, and the listening window is exactly that span, decided by a physical action rather than by the software guessing when a sentence begins or ends.
What it's being compared against
The term only makes sense next to the two other ways dictation software can decide when to listen:
- Voice-activated — the tool listens continuously and uses a pause in speech to guess that a sentence has ended, usually after some fixed number of silent seconds.
- Toggle-based — a shortcut turns listening on, and it stays on until the same shortcut (or a timeout) turns it back off.
- Push-to-talk — a key held down is the entire listening window. Nothing is heard before it goes down, and nothing is heard after it comes back up.
Where the name comes from
The term predates dictation software by decades — it's the same mechanism behind a two-way radio and the key gamers have used for voice chat since before speech-to-text dictation existed as a consumer product: press to speak, release to stop, with no state in between where the channel is open but nobody means to be transmitting.
What it actually changes about the interaction
Remember to confirm the delivery window before Friday.
Two things follow directly from listening only while the key is held. First, there's nothing to leave running by accident — a voice-activated tool that keeps listening into a phone call taken right after, or a toggle left on longer than intended, are both failure modes push-to-talk doesn't have, because there's no state to forget about. Second, the moment the key goes down is also the moment most push-to-talk tools decide which application should receive the text, before a word is even spoken.
What it trades away
Push-to-talk needs a hand free to hold the key for the length of the sentence, which is a real limitation for anyone not seated at the keyboard — carrying something, standing across a room from the Mac, or otherwise unable to press and hold anything. Voice activation's whole advantage is not needing that; it costs the guessing problem in exchange for hands-free use.
Where this shows up in an actual Mac dictation tool
The full page on push-to-talk dictation for Mac covers what this model looks like in daily use — including what it costs you against the alternative and who it actually suits. For a comparison against a tool built around continuous, always-on listening for spoken commands rather than a held key, see Coii VoiceInput against Apple's Voice Control, a different category of always-listening tool entirely.
How it fits into the broader term
Push-to-talk is one answer to a problem every piece of dictation software has to solve somehow: deciding when to start and stop listening. It isn't the only answer, and which one suits a given person depends more on whether a free hand is available than on which is objectively better.