OpenSuperWhisper for MAC

Free, open-source Whisper dictation app for Apple Silicon Macs: record with a shortcut, transcribe speech to text and drop in audio files.

OpenSuperWhisper Overview

OpenSuperWhisper is a macOS app for real-time audio transcription. Press a global keyboard shortcut, speak, and the app turns your recording into text using the Whisper speech recognition model. It offers two transcription engines, Whisper (through whisper.cpp) and Parakeet, and you download the models you want directly from within the app. You can also drag and drop audio files onto it and let it work through them in a queue.

OpenSuperWhisper is free and open source under the MIT license, made by the GitHub developer Starmel. It supports many languages with automatic language detection and has an autocorrect for Japanese, Chinese and Korean text. Version 0.1.0 adds a setting that puts a space after each sentence, opens audio files from Finder and fixes CPU use while the app is idle. It runs only on Apple Silicon Macs and needs macOS 14 Sonoma or later; Intel support is not available. On first use you grant it the permissions it asks for.

Key Features

OpenSuperWhisper turns your voice into text on the Mac.

  • Real-time transcription: record and get text right away.
  • Two engines: Whisper and Parakeet, models downloaded in the app.
  • Global shortcuts: a key combination or a single modifier key such as Fn.
  • Hold to record: hold the shortcut and release it to stop.
  • Audio files: drag and drop files, processed in a queue.
  • Microphone choice: built-in, external, Bluetooth or iPhone mics from the menu bar.
  • Many languages: with automatic language detection.
  • Asian autocorrect: tidies Japanese, Chinese and Korean text.

What’s New in Version 0.1.0

OpenSuperWhisper 0.1.0 adds new options and fixes several bugs.

  • New setting to add a space after punctuation in transcriptions
  • Cmd+, now opens Settings
  • Open audio files from Finder by double-click or Open With
  • Better handling of audio formats, including MP4 containers
  • Fixed CPU use while the app is idle
  • Fixed automatic language detection

OpenSuperWhisper FAQ

Is OpenSuperWhisper free?

Yes. OpenSuperWhisper is free and open source under the MIT license.

What does OpenSuperWhisper do?

It records audio and transcribes it to text in real time using the Whisper model. It can also transcribe audio files you drag and drop onto it.

Which transcription engines does it use?

It offers two engines, Whisper (through whisper.cpp) and Parakeet. You download the models directly from within the app.

Which macOS version does OpenSuperWhisper need?

OpenSuperWhisper 0.1.0 needs macOS 14 Sonoma or later.

Does OpenSuperWhisper run on Intel Macs?

No. It runs only on Apple Silicon Macs. Intel support is on the project's to-do list but is not available yet.

How do I start recording?

Use a global keyboard shortcut: a key combination or a single modifier key such as Left Command, Right Option or Fn. With hold-to-record, you hold the shortcut while speaking and release it to stop.

Which languages does it support?

It supports many languages with automatic language detection, plus an autocorrect for Japanese, Chinese and Korean text.

Can I choose the microphone?

Yes. You can switch between built-in, external, Bluetooth and iPhone (Continuity) microphones from the menu bar.

Can I install OpenSuperWhisper with Homebrew?

Yes. Run brew install opensuperwhisper.

Where does the OpenSuperWhisper download on opensuperwhisper.FULLMAC.app come from?

The download button leads to the official OpenSuperWhisper releases on GitHub at https://github.com/Starmel/OpenSuperWhisper/releases/latest. opensuperwhisper.FULLMAC.app does not change or repackage the app.

Similar Apps

FULLMAC.app does not host this app. Download buttons lead to the developer’s own pages.