By the time a translator starts typing the target sentence, the actual translating is done — the meaning has been carried across, the phrasing has been chosen, the tone has been matched. What's left is mechanical: getting that already-decided sentence into a document, one keystroke at a time. This page is about replacing that last mechanical step with a held key. It is worth saying plainly, up front: this app does not translate anything. It transcribes speech in whichever language is spoken, into text in that same language. The translating still happens in the translator's head, the same as it always has.
What actually costs a translator the time
Rarely the decision — a working translator usually knows the target sentence within moments of reading the source. What takes the time is typing that sentence out, on top of the sentence that's already been said in someone else's head. A page of source text becomes a page of typing even though the thinking behind it was finished paragraphs ago.
Where a held key fits into an actual translation session
Drafting the target-language sentence, said instead of typed
Hold ⌥ Space, say the sentence in the target language exactly as you'd want it written, let go — it lands wherever the cursor already is, through the accessibility API where the document editor supports it and a synthetic paste otherwise, clipboard restored either way.
Le rapport sera livré vendredi avant midi, sous réserve du délai habituel de relecture.
A note in the source language, without switching a setting
A translator working between two languages in one sitting sometimes needs a quick note back to a source-language colleague or client, not just the target-language draft. The default engine detects Chinese and English mixed inside a single utterance and punctuates the Chinese half automatically, without a separate step to switch languages first — a narrower claim than full bilingual support across any two languages, but a real one for that specific pair.
Proofreading and revision notes to a client
Explaining a translation choice — why a phrase was rendered one way instead of a more literal alternative — is prose, typed into an email or a comment thread, and benefits from being said the way any explanation does.
Working inside a CAT tool
Translation-memory software and CAT tools have their own text fields for the target segment, and dictation reaches them the same way it reaches anything else — through the accessibility API where the tool supports it, a synthetic paste otherwise, with the clipboard restored afterward either way.
What a working session's history actually holds
- 09:404.4sLe rapport sera livré vendredi avant midi, sous réserve du délai habituel de relecture.Word
- 10:153.1sDer Liefertermin bleibt Freitag, vorbehaltlich der üblichen Prüfzeit.Word
- 13:022.7sKept "delai habituel" rather than translating it more literally, matches the client's prior glossary.Email
Every dictated segment is written to local history the instant the key is released, before anything else happens to it, and that history stays on the machine it was said on. Nothing about a draft translation is uploaded anywhere to be processed, and the app works exactly the same with the network off — relevant for a translator working from a client's source files under a confidentiality agreement that rules out any tool that phones home.
What this deliberately does not do
It does not translate. Saying a sentence in one language types that sentence in that language, not a rendering of it in another. It has no real-time translation feature of any kind, and nothing on this page should be read as suggesting otherwise. A broader comparison of dictation apps across languages covers which tools in this category do publish a translation feature and which, like this one, only transcribe — worth reading before assuming any dictation tool closes that gap.
It also has no custom vocabulary, so a specialized term, a proper noun or an industry-specific phrase comes out however the engine hears it and needs the same review any dictated segment would get before it's finalized in a professional translation.
Where the published language list falls short
This app names ten languages explicitly and describes a broader engine only as covering "dozens" more, without a specific count. Wispr Flow publishes support for 100-plus languages, a materially wider list — a real gap for a translator who regularly works in a language pair outside this app's named set, and worth checking against the specific languages a given job actually requires before assuming this is the right tool for it.
Where Apple Dictation already covers a lot of this, free
Apple Dictation is already installed on every Mac, free, supports multiple languages depending on the Mac and macOS version, and lets a session switch languages manually by clicking the abbreviation near the cursor. It does not translate either. What a dedicated hotkey mainly adds over the built-in option is a trigger that starts the instant the key goes down, rather than after Apple's own short pause — worth trying first for anyone unsure whether a paid app adds enough to be worth it.
What happens if a segment gets interrupted mid-dictation
- 1Key downthe target is chosen here
- 2You speaklevel meter, over your work
- 3Key up
- 4Written to historybefore the engine is asked
- 5Engineon your Mac
- 6Typedabout half a second
A CAT tool that briefly loses focus, or a document editor that crashes mid-session, are ordinary hazards of a long translation session, and they can interrupt a dictated segment as easily as a typed one. What changes is where the words already are: written to local history the instant the key is released, before the engine that turns speech into text is even asked to produce a result, so an interrupted segment can be pasted back in from history rather than translated and said again from scratch.
Working across two machines on the same project
A licence covers three devices, each installed separately with no account and no sync between them. A translator who works from a desktop at home and a laptop on site with a client gets an independent history on each machine — nothing dictated on one shows up on the other automatically, which matters for anyone keeping a client's confidential source files off a machine they don't control.
Building the habit across a project
The benefit rarely shows up in any single segment being noticeably faster to produce than typing it would have been — it shows up across a full project, in how much of a working day is spent getting already-translated sentences onto the page rather than typing them out one keystroke at a time. A held key does not translate anything better than a translator already has; it only changes how quickly a decision that's already been made gets recorded.
What it costs and what it needs
$19 once, three devices, macOS 13 Ventura or later, Apple Silicon or Intel. The thirty-day trial runs the full app, no card and no account, long enough to carry a real project's worth of segments through it in whichever languages the work actually requires, and judge whether the mechanism holds up against a genuine deadline. The page on dictating email covers the correspondence side of the same job — client notes, delivery confirmations, questions about a source text — for anyone whose day is not purely drafting target-language segments.
Why the target sentence still deserves a careful read-back
Saying a translation aloud rather than typing it does not change the professional obligation to check it before it's delivered — a mistranscribed word, a name the engine heard wrong, or a punctuation mark applied incorrectly can slip into a dictated segment the same way a typo slips into a typed one. What changes is where the effort goes: less of it spent on the mechanical act of typing an already-decided sentence, and, in principle, more of it available for the read-back that any professional translation deserves regardless of how the first draft got onto the page.
Where dictation fits around, not instead of, a CAT tool's own features
Translation memory, terminology matching, and quality-assurance checks built into a CAT tool all continue to work exactly as they did before — this app only changes how a target segment gets typed into the tool's text field, not what the tool does with it afterward. A fuzzy match suggested by the CAT tool, accepted or edited, is still handled through the software's own interface; dictation is one way of getting a fresh segment into that field when no match applies and the sentence has to be composed from scratch.
A realistic session across two target languages
A plausible stretch: draft a paragraph of a French target document from a source already read and mentally translated, switch to a German delivery note for a different client an hour later, and dictate a short note back to a project manager in English explaining a terminology choice made in the French draft. None of it involves the app doing any translating — every sentence was already fully formed in the translator's head before the key was pressed. What changed across that session was only how quickly each already-decided sentence made it onto the page.
A hotkey that survives a CAT tool's own shortcuts
CAT tools tend to carry dense keyboard shortcuts of their own — for confirming a segment, inserting a tag, or navigating between them — and a system-wide dictation hotkey has to avoid colliding with those. The default ⌥ Space is chosen to sidestep the most common conflicts, but it is worth a quick check in System Settings, under Keyboard Shortcuts, against whichever CAT tool is already installed before relying on it for a full day of segments.
Why speed matters most against a deadline, not against the craft
Nothing about dictating a target sentence faster changes how carefully it was translated — the quality of the translation is decided entirely before the key is pressed, in the same thinking a translator would do regardless of how the sentence gets typed afterward. Where the time saved actually shows up is against a deadline: a document due by end of day leaves more room for a careful final read-through when the drafting itself took less of the day, rather than less room because typing consumed the hours that review would otherwise have used.
What this means for a rush job specifically
A same-day turnaround is where the gap between having already translated a sentence in your head and getting it onto the page matters most — there is simply less slack in the schedule to type slowly. Dictating the draft under that kind of deadline does not make the translation any more accurate than typing it would have; it changes how much of the available time gets spent on the mechanical act of typing versus on the proofreading pass a rushed job needs even more than an unhurried one.
Who this actually suits
A translator working primarily within this app's named languages, drafting target-language text that's already fully formed before it's said aloud, has a straightforward reason to try the thirty-day trial against a real assignment.
A translator who needs a language outside this app's named list, who wants an actual translation feature built into the dictation tool itself, or who needs a vocabulary that learns a client's specialized terminology, is better served reading the multilingual round-up first — this app was built to type what's said, in the language it's said in, and nothing more.