1
Open setup
Launch Acousmos from the menu bar. If setup is incomplete, open Preferences and use Run Setup Again.
User guide
Speak once. Let Acousmos write where your cursor already is.
Set up permissions once, choose your input style, then use a hotkey to dictate, translate, or edit text inside the app you are already using.
Quick start
From install to first sentence.
2
Grant permissions
Allow Microphone so Acousmos can listen, and Accessibility so it can paste the finished text into the focused app.
3
Choose a style
Use Minimal for everyday input, or Minimal-Translate when you want spoken language converted into the target language.
4
Press, speak, release
Put your cursor in any text field, press the style's hotkey, speak naturally, then stop recording. Acousmos pastes the result.
Permissions
Why macOS asks for access.
These prompts are normal for a global voice input app. Acousmos needs enough access to hear you and place text back into the active app.
Microphone
Required for recording your voice.
Accessibility
Required for reliable paste into the focused editor, browser, chat app, or notes app.
Input Monitoring
Only needed if macOS asks while using bare-key or push-to-talk hotkeys such as Fn, Right Command, or Right Option.
Modes
Input, translation, and editing styles.
Each style can have its own hotkey. The app defaults to a daily input style and a translation style, and you can change both in Preferences.
Input
Minimal
Best for everyday writing. It fixes obvious speech-recognition errors while keeping your phrasing intact.
Default hotkey: Right Command.
Input
Verbatim
No LLM pass. It pastes the transcription as produced by the selected ASR provider.
Input
Shape
Turns spoken fragments into cleaner written text by merging run-ons, removing filler, and adding punctuation.
Translation
Minimal-Translate
Translates faithfully into your target language while preserving your original structure as much as possible.
Default hotkey: Right Option.
Translation
Shape-Translate
Translates with more freedom to restructure the sentence into natural target-language output.
Edit
Edit
Use your speech as an instruction, such as "make this shorter" or "translate this paragraph".
Translation
Set the target language once.
Open Preferences, go to Language, and set the Translation target. Minimal-Translate and Shape-Translate will use that language when they turn speech into text.
en
zh
ja
es
zh-tw
Source
Speak in Chinese, Japanese, Spanish, or another source language.
Target
Acousmos uses the language code you set in Preferences.
Output
Translated text is inserted at the cursor in the active app.
Preferences
Where to change behavior.
Readiness
Check Microphone, Accessibility, ASR, LLM, and enabled hotkeys before your first real use.
Hotkeys
Assign separate keys for input, translation, or custom styles. Modifier-only hotkeys can work as tap or push-to-talk.
Language
Control source-language detection and the target language used by translation styles.
Providers
Choose local or configured cloud providers for transcription and LLM polish, depending on your privacy and quality needs.
Troubleshooting
If the first attempt does not work.
Nothing appears
Click into a text field first, then check Accessibility permission in System Settings.
No audio is captured
Check Microphone permission and make sure the selected input device is receiving sound.
The hotkey does not fire
Confirm the style has an enabled hotkey. For push-to-talk keys, approve Input Monitoring if macOS asks.
First use feels slow
Local transcription or polish models may need to download before the first run completes.
Names or terms are wrong
Add important vocabulary, recurring phrases, or replacements in Preferences so Acousmos can preserve them.
You need setup again
Open Preferences and choose Run Setup Again. The in-app Open Preferences Guide button points back to this page.
Ready
Use your voice anywhere you can type.
Acousmos runs from the menu bar, keeps a local archive on your Mac, and lets you decide how much AI processing to apply.
Download for Mac