Launch price $11.99 until 18 Nov, then $21.99. Pay once, 2 Macs
Pricinghushtype · features
76 features in 11 chapters, each with a short video or a live scene you can turn with the pointer, a finger or the arrow keys. Everything runs on your Mac. Where a part needs a Mac with permissions that we have not yet run ourselves, the text says what the app does and nothing more. See what is rough or not finished.
Chapter 1
Hold a key, speak, let go.
Hold Right Option, speak, and let go: the sentence is typed into the app you are in, at the cursor, as if you had typed it. Tap the key twice to keep listening hands-free, and press Escape to throw a dictation away.
Between hearing you and typing, fixed rules tidy what was said: fillers like um and uh go, a repeated word is said once, a take-back such as no wait or I mean replaces what came before it, spoken punctuation becomes punctuation, and numbers and units are written the way you would type them. The same words always give the same text.
Choose how much tidying you want per app: Exact keeps your words as said, Clean removes the fillers and take-backs, Polished adds a grammar pass. Live text shows the words in the recording pill while you speak; the cleaned sentence replaces it when you let go.
Quick start keeps the recogniser loaded for a few minutes after a dictation so the next one starts sooner. Keep the microphone ready is a separate, opt-in setting that also holds the last half second, so a word said as you press the key isn't cut.
It types through the clipboard (and puts your clipboard back) or through Accessibility, and never into a password field. If typing fails, the cleaned text waits under Unsent text in the menu.
Hush mode raises quiet speech to a steady level for whispering in an open office or late at night. Like every part that needs a microphone and permissions, it describes what the app does; we have not measured it on your Mac.
Hold Right Option and speak. Tap twice to keep listening hands-free. Escape cancels.
Pastes the text (then puts your clipboard back) or types through Accessibility, and always skips password fields.
Choose your own words, tidied words, or tidied words plus a grammar pass, per app.
Fillers, repeats, take-backs, spoken punctuation, capitals, numbers and units. The same words give the same text.
Say scratch that to drop the clause you just said, or undo the last dictation within a short window.
A small pill at the bottom of the screen with a live level meter. Cues are off by default and can be turned on.
Keeps the recogniser (and the accurate engine's model) loaded for a while after each dictation: for the minutes you choose, a third as long on battery, never on Low Power Mode. The microphone stays closed.
Off by default. Keeps the microphone open between dictations for the same minutes, with the orange dot showing, and holds only the last half second in memory, so a word said as you press the key isn't cut.
The words appear in the recording pill as you speak, newest words in view. The cleaned-up text replaces them when you let go.
For speaking softly: quiet speech is raised to a steady level, the room is faded down between words, and the pill shows a meter made for a whisper. In the menu and in Settings.
Chapter 2
Two ways to turn speech into words, both on this Mac, and the words you care about given to the recogniser before you speak.
Two ways to turn speech into words, both on this Mac. Apple's recogniser is always there and is the default. The accurate engine is a larger speech model you can download once from Settings › Engine; after that it runs with Wi-Fi off.
Before you speak, the words you care about are handed to the recogniser: your dictionary, snippet names and the packs you switched on, and, if you allow it, the names visible in the front window.
Open Settings › Engine, pick Accurate or Fast, and press Get. The app says what it will download before it starts; every file is checked against its published checksum, a dropped connection resumes, and you can delete the model any time.
When a model is used for the very first time, macOS prepares it for your Mac, which can take minutes for the accurate model. Until then dictation simply uses Apple's recogniser, so you never wait.
On a Mac without Apple's own chips, with too little memory, or if anything goes wrong, Apple's recogniser takes over by itself. Meetings, captions and hands-free listening always use Apple's recogniser.
The download comes from the tealmac download server; the model is the only optional network use besides your licence.
A larger speech model, in about 100 languages, that you download once from Settings › Engine; after that it runs on the Mac with Wi-Fi off. Apple's recogniser stays the default and takes over whenever the accurate engine can't run.
A smaller download for Macs with less memory. It starts quickly; it is less accurate than the accurate model.
Your dictionary, snippet names and the packs you switched on are handed to the recogniser before you speak, so names and jargon can come out right straight away.
Off by default. When a dictation starts, the visible labels of the front window (people, titles, buttons) are added to the words to listen for. Never password fields or what you are typing; nothing is kept.
Chapter 3
Plans from plain words, captions for anything, typing that fits, and a nod for yes.
Ask in your own words, like move today's screenshots from the Desktop into a folder called October, or set up for a call: Zoom left, Notes right, Focus on. The request becomes a short plan of numbered steps.
Live captions put the words of anything your Mac plays (calls, videos, podcasts) in a bar over every window, recognised on the Mac, translated if you like. Typing that fits the field reads a short stretch before the cursor and carries on a list or the code style. A nod says yes.
Use the command key or your wake phrase, or start a dictated sentence with do this. Every plan is shown before it runs, with a preview of the exact files from a real look at the folder, and waits for yes. Say put the files back to undo the last file plan.
Captions start from the menu. Choose the size, position, lines and background; translation uses macOS's own on-device translation on macOS 26.
A plan can only use a short list of safe actions and stays inside your home folder. It never types text, never runs scripts, never sends anything and never deletes for good; a request it can't do safely is refused with the reason.
Captions live in memory while they are on. Nothing is kept unless you press Save this transcript.
Ask in your own words with the command key or your wake phrase, or start with do this while dictating. The plan is shown as numbered steps and waits for your yes. Common requests work on any Mac; anything else uses the Mac's own on-device model on macOS 26.
Find files in your own folders by kind and date (today's screenshots, PDFs in Downloads, installers older than 30 days), then make a folder, move, copy, rename or move them to the Trash. Never deleted for good, never outside your home folder.
Set up for a call: Zoom left, Notes right, Focus on. Every part has to be understood, or nothing runs.
Before the yes, the panel shows which files a plan would touch, from a real look at the folder. Say put the files back to undo the last file plan.
Captions for whatever your Mac plays (videos, calls, podcasts), in a bar on top of every window and full-screen video. Recognised on the Mac. Size, position, background and number of lines are yours.
Translate captions into your language with macOS's own on-device translation (macOS 26), with the original underneath if you like. A language macOS hasn't installed needs its own download first.
Captions live in memory while they are on. Save this transcript writes a Markdown file with times, where you choose.
Reads a short stretch before the cursor (never a password field, never kept) and carries on a list, writes the opening word in lower case mid-sentence, and gets the spacing and full stops right.
In code, a spoken name follows the style the file already uses (camelCase, snake_case, kebab-case) and new lines keep the indentation.
Cleanup follows the language the field is already written in, among the languages you use. The recogniser still listens in its own language.
With AirPods or Beats that track head movement: nod for yes or shake your head for cancel when hushtype asks first. Read only while the question is on screen; off by default.
Chapter 4
Things people dictate every day, written the way you would type them.
Things people dictate every day come out the way you would type them: john at example dot com becomes john@example.com, five five five one two three four becomes 555-1234, and a spoken web address becomes a link you can paste.
Voice commands, lists and edits are part of dictation: new line, new paragraph, new bullet, heading two, bold and end bold, delete last word, select last sentence, capitalise that.
Just say them. A command only counts when the whole sentence is a command, so dictating a sentence that happens to contain the word copy still types it.
Open an app with open Mail or switch to Safari. Names are matched by sound, and a sentence like go to the store is still typed.
Smart formatting and lists are English only for now. You can switch smart formatting off to keep plain digits.
Emails, web addresses, phone numbers, dates and dollars-and-cents are written as data, not as words. English only.
New line, new paragraph, cap that, select all, copy, cut, paste, arrow keys, delete last word. A whole sentence only.
New bullet, new numbered item, heading two, block quote, bold and end bold. Markdown or plain, by style. English only.
Delete last word, select last sentence, capitalise that, correct a word, press enter, tab or escape, chained with and or then.
Say open Mail or switch to Safari, with the command key or as a whole sentence while dictating. Names are matched by sound, and a sentence like go to the store is still typed.
Chapter 5
The whole Mac by voice: no keyboard, no mouse. Every command and the words to say are on the voice control page.
The whole Mac without a keyboard or mouse: open and switch apps, move windows to halves, quarters or thirds, click any button by its name, show numbers on everything clickable and say one, zoom a grid down to a point, press any key or shortcut, move and select text, open any of 45 System Settings panes, read notifications and run Shortcuts.
Show text numbers puts a number on every visible word in a text field so you can select, delete or replace by number, and correct that lists the recogniser's other guesses for what you just said.
Hold a command key and say it, say a whole command while dictating, or switch on always-listening and start with your wake phrase. Say what can I say for the commands that work in the app in front.
App packs add words for browsers, Finder, Mail, Notes, Messages, terminals, code editors, Slack, Zoom, Teams and Music. Your own commands run steps written the way you say them.
Restart, shut down, log out, empty the Trash, send, delete and leave always ask first. A command sent by a link always shows what it will do and waits for your yes.
Clicking, overlays and window moves use Accessibility and read button and menu names only when you give a command. Some web and Electron fields expose little to Accessibility; there, commands that need word positions say so.
Hold a command key and say it, say a whole command while dictating, or switch on always-listening and start with your wake phrase. Go to sleep and wake up pause it.
Open, switch, quit, hide and reopen apps by name, open folders and recent files, open websites and search the web in your browser.
Halves, quarters, thirds, centre, fill, bigger or smaller, next display, minimise, full screen, tile two apps, and put the window back.
Click by name, show numbers and say one, or zoom a 3 by 3 grid down to a point. Double, right click, hold, drag, hover and scroll.
Press any key or shortcut by name, type text literally, spell with the phonetic alphabet, and say camel case or snake case.
Move by word, line, sentence or paragraph, select a phrase or from one phrase through another, replace, insert before or after, delete and bold.
Show text numbers to select, delete, capitalise, replace or insert by word number. Correct that lists the other guesses for what you just said; choose 2, or spell it.
Volume, brightness, music, dark mode, lock, screenshots, Spaces, Mission Control, Spotlight, Control Centre, Wi-Fi, battery and time. VoiceOver on and off.
Open any of 45 System Settings panes by name, from Wi-Fi to Privacy and Security.
Read the notifications on screen, open one by number, dismiss one or clear them all.
Words for browsers, Finder, Mail, Notes, Messages, terminals, code editors, Slack, Zoom, Teams and Music. Packs are JSON you can share.
A phrase runs steps written the way you say them, everywhere or in one app. Run any Shortcut by name.
Restart, shut down, log out, empty the Trash, send, delete and leave always ask; answer yes or cancel by voice or click.
Lists the commands for the app in front first. A large command panel, a sound and spoken answers are options.
Chapter 6
The right finish for where your cursor is.
The right finish for where your cursor is: formal in Mail, casual in Messages, lower case and symbols in Terminal and code editors. Every style can be changed.
For people who type code: say dash dash all and get --all, say camel case user name and get userName, say a path and get an @ reference. The Prompt style sorts a long spoken request for a coding agent into the goal, then Requirements, Constraints and Done when.
Styles follow the app in front automatically, or pick one from the menu. Rules can switch dictation off in chosen apps, during calls, in quiet hours, or by Focus.
Password fields are always skipped, and the do-not-dictate list keeps it out of apps you name.
Formal in Mail, casual in Messages, lower case and symbols in Terminal and code editors. You can change any of them.
Quiet hours, off in chosen apps, a style per app, off during calls, and a Focus filter that pauses it or picks a style.
Say symbols, paths, flags and casing: dash dash all becomes --all. Camel, pascal, snake, kebab and constant case by voice.
A long spoken request is sorted into the goal, then Requirements, Constraints and Done when. Spoken paths become @ references; bullets, inline code and code blocks marked with a language come by voice. Short requests stay as said.
Name the apps where it should never type. Password fields are always skipped.
Chapter 7
Spelled your way, and yours to take with you.
Names and terms come out as you spell them. Say a short trigger and get a longer text. Templates such as commit message, pull request description, meeting recap and stand up fill in the date, time, weekday, clipboard and app.
Teach it: load your last dictation, fix it, and it proposes dictionary entries for the words it heard wrong. Nothing is learned until you press Add.
Bring your words from another dictation app as CSV, TSV, JSON, a plain list or a -> b lines, and export them the same way. Software, legal and health packs add common terms.
Your dictionary is also handed to the recogniser before you speak, so the right spelling can come out straight away.
Names and terms come out as you spell them. Say a short trigger and get a longer text.
Import a word list from another dictation app as CSV, TSV, JSON, a plain list or a -> b lines, and export the same way.
Load your last dictation, fix it, and it proposes dictionary entries for the mishearings. Nothing is learned until you press Add.
Say a name to get a template: commit message, pull request description, meeting recap, support reply, stand up, bug report. Fills in date, time, weekday, clipboard and app.
Software, legal and health word packs, phonetic hints on any word, per-project dictionaries, and packs you can import and export.
Chapter 8
Talk to text you already wrote.
Talk to text you already wrote: select it, hold the key and say make it shorter, more formal or fix the grammar. Read-back reads the selection or your last dictation aloud so you can proof by ear.
Mechanical edits are done by rules. Open-ended ones use the Mac's own on-device language model where macOS has one; without it the app says it can't, rather than guessing.
Nothing is sent anywhere: the edit happens on the Mac, and you see the result before it replaces your selection.
Select text, hold the key, say make it shorter, more formal or fix the grammar. Rules do mechanical edits; the Mac's on-device language model does open-ended ones where it has one.
Reads the selection or your last dictation aloud with the system voice, so you can proof by ear. From the menu.
Chapter 9
Recordings in, Markdown out.
Meeting mode records your microphone and the sound your Mac plays and keeps a transcript with You and Others turns, timestamps and your typed notes. Decisions and action items are picked out by phrase rules.
File transcription takes audio or video you drop in and saves .txt, Markdown, .srt or .vtt. A watch folder transcribes whatever lands in it. The voice inbox saves a spoken thought as a dated line in a daily note.
Start meeting notes from the menu. Export as Markdown with speakers and times, turn action items into Reminders, and make Calendar events from sentences with a time, only when you press the button.
The other side of a call needs macOS's Screen Recording permission; the app records sound only and never the picture. Tell people they are being recorded where the law or courtesy requires it.
Records your microphone and the sound your Mac plays, and keeps a transcript with You and Others turns, timestamps and notes.
Markdown with speakers and timestamps, optional Obsidian-style frontmatter, and safe file names. Decisions and action items come from phrase rules.
Turns action items into Reminders with due dates, makes Calendar events from sentences with a time, and drafts a follow-up email, only when you press the button.
Chapters from pauses, a short extractive summary, the questions asked out loud, terms to review, and Markdown export.
Drop audio or video to queue it. Save as .txt, Markdown, .srt or .vtt. A watch folder transcribes whatever lands in it.
Say a thought and it is saved as a dated line with optional tags in a daily note or one file.
Chapter 10
No hidden numbers, no lost text.
Wi-Fi on. Speech is still recognised on this Mac.
Recognition and cleanup run on the Mac with network recognition off. No account, no analytics, no screenshots, and audio is never written to disk. Turn Wi-Fi off and it keeps typing, as long as the language's speech model is on the Mac.
Stats are counts only: words, dictations, streaks, busiest hour, per-app counts, your own timings, and a time-saved figure with its working shown.
Keep history or not, delete it after 1, 7, 30 or 90 days, switch it off for single apps, export everything in one click, or delete everything in one click. Your licence stays.
The app goes online only for the licence server and, if you choose it, the one-time download of the accurate engine. It never sends what you say or write.
Recognition and cleanup run on the Mac with network recognition off. No account and no screenshots. Voice commands read button and menu names only when you give one. The optional accurate engine is downloaded once and then runs offline.
Cleaned text is saved before it is typed. If typing fails or the app stops, it waits in the menu under Unsent text.
Searchable, on this Mac, off if you like, auto-deleted after 1, 7, 30 or 90 days, and off for single apps.
One click exports everything. One click deletes history, meetings, stats and transcripts. Your licence stays.
Words, dictations, streaks, busiest hour and per-app counts. No words or audio are kept for them.
Words divided by your typing speed, minus speaking time, never below zero. The last 7 days use the same rule.
Shows your own median and slowest 1 in 20 time from key release to typed text, over your last 200 dictations. Timings only.
Words per minute, fillers per 100 words, the trend against last week, and a pace graph. Counts only.
Chapter 11
More than one language, and no clicking.
The menus and windows come in 21 languages, and cleanup has its own rules for English, German, French, Spanish, Portuguese and Hindi, switching between your languages sentence by sentence.
hushtype:// links, Shortcuts actions and a small command-line tool start, stop and steer dictation from your own scripts.
Set the recogniser's language for each app. Links that would turn the microphone on ask first unless you allow your own scripts in Settings › Commands.
The interface translations are machine-made and not yet checked by native speakers. The recogniser listens in one language at a time.
The app's menus and windows are translated. The translations are machine-made and not yet checked by native speakers.
English, German, French, Spanish, Portuguese and Hindi have their own fillers, spoken punctuation and take-backs. Switches between your languages sentence by sentence.
Set the language the recogniser listens in for each app. It listens in one language at a time.
hushtype:// links start, stop, toggle, cancel and undo, switch a mode, start a meeting, listen for a command or run one. A command sent by a link always shows what it will do and waits for your yes, because a web page could open a link too; a link can never answer that question itself. Shortcuts actions and a small command-line tool use the same links.
One licence key, no subscription, works on 2 Macs. No account and no sign-in.
Exact behaviour of the newest features.
The same mistake twice is what people dislike most about dictation. Open Teach it from the menu, load your last dictation, fix the text and press Find changes. It proposes “when it hears X, write Y”, or “always write Y”.
Spoken data is written as data, not as words. It is in Settings › General, on by default, and off in Exact mode and Code style. English only, by rules that can be wrong at the edges.
Dollars and cents are written as dollar amounts. Web endings it knows include com, org, net, io, dev, edu, gov, app, info, biz, xyz, tech, uk, de, fr, ca, au and co. A spoken run of 7 digits now becomes 555-1234; switch smart formatting off for plain digits.
Say a command as a whole sentence and it acts instead of typing. A command inside a longer sentence stays words. Off per app in Settings › General.
“Open Mail” and “switch to Safari” work as a whole sentence while dictating, or with the command key; the voice control page has every command. Sending key presses and opening apps are written and build, but we have not yet run them ourselves in every app.
Say the name on its own and the template is typed, with the date, time, weekday, clipboard and current app filled in where it uses them.
Standup, bug report and meeting notes were already there; the Add built-in templates button adds the rest. Macro packs import and export as JSON so a team can share them.
Cleaned text is saved as a draft before it is typed. If typing fails, or the app stops, the text waits in the menu under Unsent text, with Copy and Delete.
The Stats window shows your own timings: the median and the slowest 1 in 20 time from key release to typed text, over your last 200 dictations. Timings only; no text is kept for it.
It is a way for you to check whether it feels fast on your Mac. We do not publish speed numbers, because yours depend on your Mac.
No hidden numbers. hushtype shows how it got the figure, and the last 7 days use the same rule.
time saved = words ÷ your typing speed − time you spent speaking (never below 0)
Example with made-up numbers: 300 words at a typing speed of 40 words a minute is 7.5 minutes of typing. Speaking them took 2 minutes. Time saved: 5.5 minutes. You set your own typing speed in Settings.
Launch price: $11.99 until 18 Nov, then $21.99. One-time, 2 Macs, no subscription.
Buy, then download. No account, no subscription, no word limits.