Skip to content

Legal dictation is one of the oldest use cases in the category, and for decades the answer was a physical dictaphone, a cassette or a digital file handed to a typist, and a chain of custody to think about along the way. A Mac dictation key replaces a narrower slice of that than it might first sound — it does not transcribe a recording, and it has none of the practice-management structure a dedicated legal tool builds around dictation. What it replaces is the keyboard: memos, correspondence and billing narratives typed by voice, straight into the document, by the person who dictated them.

What actually slows a lawyer down at a keyboard

Volume, mostly. A single matter can generate dozens of short documents in a day — an email to opposing counsel, a note to file, a paragraph added to a draft agreement, a line of billing narrative recorded before the memory of what the hour was spent on fades. None of that is complex writing. It is just a lot of it, typed between other things, and every one of those documents is also something that should never leave the machine it was written on without a deliberate decision to send it somewhere.

Correspondence, dictated at the moment the position is clear

Hold ⌥ Space, dictate the paragraph, let go — the words land in whatever draft was open, through the accessibility API where the document software supports it and a synthetic paste otherwise, clipboard restored either way.

⌥ Space→ Pages8.3s

The proposed amendment to clause four changes the renewal term from automatic to opt-in, and we would recommend flagging that shift explicitly before the client signs.

A memo paragraph dictated once, at the moment the position is clear, rather than reconstructed at the keyboard afterward.

Billing narratives, written the same day the time was spent

A line of billing narrative dictated immediately after a call or a research session tends to be more specific and more defensible than one reconstructed from memory at the end of the week — the difference between "reviewed correspondence" and an actual account of what was reviewed and why. Speaking it while it is still fresh costs seconds; typing it accurately three days later often does not happen at all.

Notes to file, without opening a separate window

A note to file is exactly the kind of short, immediate documentation that gets skipped when it requires switching to a different application first. A held key removes that step: the note goes into whatever document or case-management field is already open, the instant a call or a meeting ends.

Why nothing leaving the machine matters more here than most places

Privilege and confidentiality obligations do not pause for the convenience of a cloud dictation tool, and "the vendor has a privacy policy" is a promise about how a company behaves with data it has already received — not a guarantee that the data was never sent. Nothing here is uploaded at any point, on any tier: no audio and no transcript leave the machine, and the app works with the network off entirely. That is an architectural fact a security review can verify, not a policy it has to trust. A round-up of dictation tools that do not train on what you say and a broader look at private, local speech-to-text on Mac are worth reading alongside this page if that property is the whole reason you are here.

Where this has no equivalent to Dragon

Said plainly, because the gap is real and worth knowing before relying on this for a full legal-dictation workflow: Dragon built custom voice commands to insert standard boilerplate and macros to automate multi-step workflows — capabilities this app simply does not have. It types what you say and nothing else. It also has no custom vocabulary to train on case names, party names or citation formats; an unusual name comes out the same way every time and has to be corrected by hand, same as any general dictation tool. For voice commands and macros specifically, Talon is the tool built for that job, and it is worth reading about honestly rather than expecting this app to grow into it.

What a working afternoon actually looks like

Dictate the memo paragraph the moment the position on a clause is settled. Dictate the billing narrative right after the call ends, while the specifics are still easy to state precisely. Dictate the note to file before moving to the next matter, rather than trying to reconstruct it at the end of the day from a half-remembered sequence of calls. None of it required opening a recorder, naming a file, or handing anything to be transcribed later — the words were typed where they belonged, in the moment they were said.

  1. 1Key downthe target is chosen here
  2. 2You speaklevel meter, over your work
  3. 3Key up
  4. 4Written to historybefore the engine is asked
  5. 5Engineon your Mac
  6. 6Typedabout half a second
The words exist in local history before the transcription engine is even asked to produce them — a slow response or a lost paste never costs the sentence.

What this does not replace: a recorded dictaphone workflow

The older model of legal dictation — speak into a recorder, hand the file to a typist, get a draft back — is a different job from the one this app does, and it is worth being precise about the difference before assuming one simply replaces the other. This app has no way to accept an audio file and transcribe it; it only handles speech as it happens, typed directly by the person speaking into whatever document is already open. Anyone whose workflow genuinely depends on handing off a recording for someone else to transcribe is not looking at a gap this app closes — it is a live, one-person tool, not a recording-and-transcription pipeline.

Matter separation across devices

A licence covers three devices, installed separately with no account and no sync tying them together — nothing about one installation knows what was dictated on another. In practice that tends to mean a desktop at the office and a laptop carried between locations, each with its own local history and neither one syncing a word to the other, let alone to a server. For anyone weighing conflicts or ethical-wall concerns about where dictated material lives, the honest answer is that it stays exactly on the machine it was dictated on, and moving it anywhere else is a deliberate, manual step rather than something the app does on its own.

Choosing a hotkey that will not collide with case-management software

Case-management and document-assembly tools often claim their own set of keyboard shortcuts, and the most common reason a system-wide dictation key appears not to work is another application already bound to the same combination. The default ⌥ Space is chosen to avoid the obvious collisions — nothing types a literal space while holding Option — but it is worth a few minutes on the first day checking System Settings, under Keyboard Shortcuts, against whatever practice software is already installed, rather than discovering the conflict mid-dictation on a filing deadline.

What migrating a dictation habit across a firm actually costs

Moving even a single attorney's dictation habit onto a new tool has a real cost that is worth naming rather than glossing over: any standard boilerplate previously triggered by a voice command has to be rebuilt as a text snippet in whatever expansion tool the firm already uses, since there is no equivalent command layer here to carry it across automatically. Any trained vocabulary of case names and recurring terms built up over years with a specialised tool does not transfer either — the first weeks with a general dictation tool will produce more manual corrections than the system it replaced, and that adjustment period is worth budgeting for honestly rather than promising it away.

What it costs and what it runs on

$19 once, three devices, macOS 13 Ventura or later, Apple Silicon or Intel. The thirty-day trial runs the complete app, no card and no account, long enough to carry a real week of correspondence and billing narratives through it and judge the habit on actual matters rather than a demo paragraph. The free option already on the Mac is worth trying first if you have not, and the general case for a third-party app covers where that free option tends to fall short for anyone dictating all day.

What happens if a dictation is interrupted

A call that ends abruptly, a window that loses focus, or a moment where the engine responds slowly used to mean a memo reconstructed from a half-formed memory of what was about to be said. Here, the words are written to local history the instant the key is released — before the transcription engine is even asked to produce a result — so an interruption of that kind costs a copy-and-paste from history rather than the substance of what was dictated. On a day with back-to-back calls and no time to double-check every memo as it is written, that property is worth more than it sounds until the day it actually matters.

What a full week actually looks like with it running

The value rarely comes from any single memo being faster to produce — it comes from how much correspondence, how many notes to file, and how much billing narrative actually get written the same day the work happened, rather than reconstructed later from memory or left undone. A held key does not make any individual document better than a typed one would be; it lowers the friction of writing it at all, at the moment it is easiest to write accurately, which over a full week of matters is where the actual time comes back.

Who this actually suits

If most of what you dictate is correspondence, memos and billing narratives — prose, not commands — and confidentiality is a real professional obligation rather than a preference, thirty days of drafting by voice is enough to know whether it changes how much gets written and how promptly the billing gets done.

If your prior dictation workflow leaned on macros, standard boilerplate inserted by voice, or a trained vocabulary of case names that got more accurate the longer you used it, be honest about that gap before switching — none of it exists here, and a page that pretended otherwise would cost you a worse afternoon than the one it promised to fix.

The way to actually find out which of these describes you is to run a real week of correspondence and notes through it, not a single test paragraph. Dictate the next memo, the next note to file, the next billing narrative, as they come up over an ordinary week on real matters, and judge it against whether the habit is starting to feel automatic by the end of that week or still feels like friction every time the key is pressed. That difference is usually clear well before the trial runs out, and it is a more honest answer than any comparison page, this one included, can give in advance.

Questions

Is this app HIPAA or privilege compliant?
It makes no compliance certification of any kind. What it does is architectural: no audio or transcript is ever uploaded, on any tier, so there is nothing for a security review to flag about where privileged words went.
Does it learn case names, party names or citation formats?
No. There is no custom vocabulary to train — an unusual case name or citation comes out the same way every time and has to be corrected by hand, the same as any general dictation tool.
Can I dictate voice commands to insert standard boilerplate?
No. This app types what you say and does nothing else — no commands, no macros, no templates triggered by voice. Dragon built that layer; this one does not have an equivalent.
Does a dictated memo work in the document software I already use?
Yes. It works system-wide, through the accessibility API where the app supports it and a synthetic paste otherwise, so a word processor, a plain document and a case-management text field all receive it the same way.