How to use TypeWhisper on a Mac: install it, pick an engine, and dictate
By Ryleigh Newman · Published
To use TypeWhisper on a Mac, install it with Homebrew or from the DMG on GitHub, grant Microphone and Accessibility access, pick a speech engine, download a local model if the engine needs one, and press the global hotkey. TypeWhisper is open source under GPL v3 and needs macOS 14 or later.
I built YapToText, a dictation app for the Mac, so read what I say about TypeWhisper with that in mind. Every fact about TypeWhisper here comes from its README on GitHub, for version 1.7.0, read on October 11, 2026.
This app is completely free. You don't have to make an account. There are no subscriptions, no difficult downloads, and nothing is locked. Made for Apple silicon Macs. Source on GitHub.
On this page: 11 sections
Install TypeWhisper
With Homebrew:
brew install --cask typewhisper/tap/typewhisperOr download the latest DMG from its GitHub Releases page. Then follow the README's quick start:
- Open Settings and grant Microphone and Accessibility access. Accessibility is under System Settings, Privacy and Security.
- Pick an engine. If it is a local one, download a model.
- Trigger the global hotkey and say something.
If TypeWhisper says no model is loaded, step two is the one that was skipped: a local engine has nothing to run until its model has downloaded.
What your Mac needs
- macOS. 14.0 Sonoma or later.
- Chip. Apple silicon is recommended. The app is a universal binary, so it also runs on Intel Macs.
- Memory. 8 GB at least, and 16 GB or more for larger models.
- Newer macOS for some features. Apple Translate needs macOS 15 or later. Apple Intelligence and Apple's SpeechAnalyzer engine need macOS 26 or later.
From TypeWhisper's README, read October 11, 2026.
| Memory | Models the README recommends |
|---|---|
| Under 8 GB | Whisper Tiny, Whisper Base |
| 8 to 16 GB | Whisper Small, Whisper Large v3 Turbo, Parakeet TDT v3, Voxtral Mini 4B |
| Over 16 GB | Whisper Large v3 |
Dictating
- Three ways to trigger. Push to talk, toggle, or a hybrid of the two, from a global hotkey. A single modifier key such as Command or Option can be the hotkey.
- Where the text goes. It is pasted into the app in front.
- Undo. Undo Last Dictation removes the last inserted dictation, and Restore Raw Transcript swaps it for the unprocessed text. Both are in the menu bar and can have hotkeys.
- Cancel. Double Escape, single Escape, or immediately, or cancellation can be turned off.
- Recovery. Failed recordings and the last three successful dictations are kept for up to 24 hours so they can be retried.
Engines, local or cloud
TypeWhisper is a toolbox of speech engines. Local ones include WhisperKit, Parakeet and Apple's SpeechAnalyzer, and the README says a local engine keeps audio on your Mac. Cloud engines work only after you set up an API for one. A live preview of the text while you speak comes with WhisperKit. Compatible local models can also be imported from a Hugging Face repository or a folder.
Workflows
Workflows are how TypeWhisper changes its behavior by context. You create them in Settings, Workflows, and a workflow can set the language, the engine, a prompt and whether to submit automatically. The README's examples: German with Whisper Large v3 in Mail, English with Parakeet in Slack, and automatic submit in Terminal.
- An app and website match together wins first.
- Then a website match in any browser.
- Then an app match.
- Then the Always workflow, as the fallback.
Prompts that rewrite text need a language model: Apple Intelligence on macOS 26 or later, a cloud provider you add, or the Local LLM plugin on Apple silicon.
Files, subtitles and the command line
- Files. Drop several audio or video files to transcribe them as a batch, or start from Finder.
- Subtitles. Export a transcription as SRT or WebVTT with timestamps.
- Dictionary and snippets. Terms, corrections, and text shortcuts with placeholders such as
{{DATE}}. Vocabulary can be imported from Wispr Flow, Handy or a CSV file. - Command line. Install the tool from Settings, Advanced, CLI Tool. It needs the local API server running.
typewhisper status
typewhisper transcribe recording.wav --language de --json
typewhisper export settings.jsonWhat is free and what is Premium
The README lists these under Premium: meeting automation with calendars, correction learning, iCloud sync of History and Inbox between Macs, and Cloud Folder Sync for the dictionary and snippets. Dictation, engines, workflows, file transcription and subtitle export are listed outside that section. The TypeWhisper alternative page has the prices, and the Windows, iOS and Android versions its site lists.
Uninstall TypeWhisper completely
brew uninstall --cask typewhisper removes the app, but the README says it does not reliably remove everything else. Its manual steps, after quitting the app:
rm -rf /Applications/TypeWhisper.app
rm -rf ~/Library/Application\ Support/TypeWhisper
rm -f ~/Library/Preferences/com.typewhisper.mac.plist
rm -rf ~/Library/Group\ Containers/2D8ALY3LCL.com.typewhisper.mac
rm -f /usr/local/bin/typewhisperThen open Keychain Access, search for com.typewhisper.mac.apikey and delete the matching items, which hold API keys and the license. The last line above applies only if you installed the command line tool. Recordings you exported live in Documents, TypeWhisper Recordings, and stay until you delete them.
TypeWhisper and YapToText
Both are GPL-licensed Mac dictation apps with a free core. TypeWhisper gives you many engines, workflows, batch files and subtitles, and asks you to assemble the setup. YapToText is the opposite trade: one speech model and one cleanup model, both inside the App Store download, working from the first launch, on Apple silicon only. TypeWhisper vs Handy and superwhisper vs TypeWhisper set it against other apps.
This app is completely free. You don't have to make an account. There are no subscriptions, no difficult downloads, and nothing is locked. Made for Apple silicon Macs. Source on GitHub.
The short version
- Install. Homebrew or the DMG from GitHub.
- Set up. Microphone, Accessibility, an engine, and a model.
- Needs. macOS 14 or later, 8 GB of memory.
- Uninstall. Delete the app, its Application Support folder, preferences, group container and Keychain items.
If TypeWhisper changes these steps, open an issue on GitHub and I will update this page. Every one gets read.
This app is completely free. You don't have to make an account. There are no subscriptions, no difficult downloads, and nothing is locked. Made for Apple silicon Macs. Source on GitHub.
Related
Sources
- https://github.com/TypeWhisper/typewhisper-mac (checked 2026-10-11)
- https://www.typewhisper.com/en/ (checked 2026-10-11)
- https://github.com/ryleighnewman/YapToText (checked 2026-10-11)
- https://apps.apple.com/us/app/yaptotext/id6786382289 (checked 2026-10-11)
Questions
How do I install TypeWhisper on a Mac?
Run brew install --cask typewhisper/tap/typewhisper, or download the DMG from its GitHub Releases page. Then grant Microphone and Accessibility access and pick an engine.
Why does TypeWhisper say no model is loaded?
A local engine needs its model downloaded first. The README's quick start has you pick an engine and download a local model before the first dictation.
Is TypeWhisper safe?
It is open source under GPL v3, and its README says audio stays on your Mac with a local engine. Cloud engines are used only if you set one up.
Does TypeWhisper run on Intel Macs?
Yes. The README calls it a universal binary that runs on Apple silicon and Intel, with Apple silicon recommended.
Which TypeWhisper model is best for my Mac?
The README recommends Whisper Tiny or Base under 8 GB of memory, Whisper Small, Large v3 Turbo or Parakeet TDT v3 for 8 to 16 GB, and Whisper Large v3 above 16 GB.