Stay in the flow
Capture a thought without opening a separate recorder, moving text between apps, or interrupting your train of thought.
RegardingWork Voice
Your voice, directly into your work.
Private, on-device push-to-talk dictation in English or Thai. Click where you want to write, hold fn while you speak, then release to insert your words.
Free for Apple silicon Macs running macOS 14 or later. Signed with Developer ID and notarized by Apple.
Built for privacy by default
RegardingWork Dictate is push-to-talk dictation for people who want to stay in their work—without sending speech to a transcription cloud.
On-device transcription
Dictation is processed locally after the speech model is prepared.
No cloud transcription
Audio and transcripts are not sent to a RegardingWork server or third-party transcription service.
No transcript history
Your words are inserted at the cursor without retaining a transcript archive.
No telemetry
The app does not transmit usage analytics or diagnostic events.
How it works
One gesture. Three steps. No separate recorder or copy-paste workflow.
-
Click where you want to write
Place the cursor in a standard text field in the app where you are working.
-
Hold and speak
Hold the fn or Globe key while you talk. The app listens only during the push-to-talk interaction.
-
Release to insert
Release the key. Dictate transcribes on your Mac and inserts the result at the cursor.
Dictate in English or Thai
Choose the language you want from the waveform icon in the menu bar. Your selection is remembered for future launches.
English
Select English for everyday messages, notes, drafts, and project updates.
Thai
Select Thai to prepare the multilingual model. The first setup is an approximately 1.6 GB download; transcription then runs locally.
Why people use Dictate
Use one simple gesture
There is no recording dashboard to manage. Hold to talk and release when you are finished.
Keep sensitive work local
Microphone audio is held in memory during dictation, transcription runs locally, and the app does not create a transcript history.
Know when the app is ready
A visible setup window guides permissions and model preparation. Once ready, the waveform icon in the menu bar shows Dictate is running.
Start automatically
Enable Start RegardingWork Dictate at Login from the menu so push-to-talk is ready after restarting your Mac.
Write across your Mac
Use dictation in standard editable text fields across common Mac applications. Secure fields and apps that restrict automated input may not accept insertion.
Suggested use cases
- Draft messages and email replies
- Capture notes while reviewing documents
- Turn rough thoughts into a first draft
- Write status updates and project summaries
- Add comments without leaving the current application
- Reduce repetitive typing during a busy workday
Your words stay yours.
RegardingWork Dictate is built around a local-first data model.
It does not send microphone audio, transcripts, telemetry, or usage events to RegardingWork. Dictated text is inserted into the active application and is not retained by RegardingWork Dictate.
The app needs network access during initial setup to obtain its on-device speech model and may use the network when you intentionally download software updates. Dictation processing itself runs on the Mac after setup.
Note: after text is inserted, the destination app’s own privacy and sync policies apply.
Product requirements
- Apple silicon Mac
- macOS 14 or later
- Microphone permission
- Accessibility permission for push-to-talk and text insertion
- Internet connection during initial speech-model setup
- Approximately 1.6 GB of model storage when using Thai
FAQ
Does my audio leave my Mac?
No. Microphone audio is processed locally and is not transmitted to RegardingWork or a cloud transcription provider.
Does the app save what I dictate?
No. RegardingWork Dictate does not maintain a transcript history. The transcribed text is inserted into the application you are using, which may have its own storage, synchronization, or privacy behavior.
Is an internet connection required?
An internet connection is needed during initial setup to download the selected on-device speech model. The Thai model is approximately 1.6 GB. After the model is prepared, dictation processing runs locally. Intentionally downloading app updates also requires a connection.
Which languages are supported?
Version 0.1.4 supports English and Thai. Choose your language from the waveform menu-bar icon. Selecting Thai prepares the multilingual on-device speech model the first time you use it.
Why does it need Microphone and Accessibility access?
Microphone access lets the app hear you while you hold the push-to-talk key. Accessibility access lets it detect that key globally and insert the finished text at your cursor.
Where can I dictate?
RegardingWork Dictate is designed for standard editable text fields in common Mac applications. Password fields, secure inputs, and some applications that restrict automated input may not accept inserted text.
Which Macs are supported?
RegardingWork Dictate supports Apple silicon Macs running macOS 14 or later.
How do I know it is running?
The setup window confirms when the app is ready, and a waveform icon appears in the Mac menu bar. The menu also provides access to setup and diagnostics.
Can Dictate start automatically when I restart my Mac?
Yes. Open the waveform menu and enable Start RegardingWork Dictate at Login.
Say it while it is fresh.
Spend less time moving between thought and text. Download RegardingWork Dictate and dictate privately on your Mac.
Download v0.1.4 DMGOpen-source and free to download. Version 0.1.4 is signed and notarized for Apple silicon Macs. View the release details and checksums, then complete Microphone and Accessibility permissions on first run.
Part of RegardingWork Voice
Dictate inserts speech at your cursor. Need a local record of full meetings instead?