Repolode
cmagnussen

cmagnussen/blitztext-app

Local-first macOS menu bar dictation app that transcribes speech with Parakeet or Whisper on your Mac and pastes it at the cursor.

Open-source alternative to:Apple DictationOtter.ai

Stars
265
Forks
99
Open issues
7
Last commit
2 days ago

What is it?

Blitztext is a dictation app for macOS. You hold a keyboard shortcut, speak, and release it; the speech is transcribed on your Mac and the text is inserted where your cursor is. Speech recognition uses Parakeet Ultra or OpenAI Whisper via WhisperKit, both running locally with Core ML on the Neural Engine. The app has no account or backend, and audio does not leave the device.

Beyond plain transcription, four shortcut-selected modes rewrite speech into clean written text, a calm message, or text with emojis. These modes use Apple Intelligence on macOS 26, or an optional user-supplied provider such as OpenAI, Anthropic, Gemini, Mistral, OpenRouter, Ollama or LM Studio. The app also includes a beta meeting recorder that produces speaker-separated transcripts and summaries, plus audio file transcription.

The project is written in SwiftUI and AppKit, is MIT licensed, and requires macOS 14 or newer, with Parakeet needing Apple silicon.

Key features

  • 01Hold a shortcut to dictate, with text inserted at the cursor and the clipboard restored afterwards.
  • 02Local transcription with Parakeet Ultra or Whisper, with automatic language detection.
  • 03Four writing modes for exact text, clean text, calm messages, or emoji text.
  • 04Per-app and per-website mode settings.
  • 05Optional meeting recording with speaker separation and summaries (beta).
  • 06Transcribes audio files such as voice memos and WhatsApp voice messages.

Use cases

  • Dictating emails, chat messages, and documents into any macOS app without sending audio to a cloud service.
  • Turning rough spoken notes into cleaned-up written text.
  • Transcribing in-person meetings on a Mac with speaker labels and a summary.
  • Converting voice memos into text on a Mac.

Quick start

  1. 1

    Download the ZIP from the latest release page

  2. 2

    Move Blitztext.app to Applications, open it, and allow Microphone and Accessibility

  3. 3

    In the settings (Modelle), load a local speech model once, then hold fn + Shift, speak and let go

Things to know

  • Requires macOS 14 or newer, and Parakeet needs Apple silicon.
  • The writing modes using on-device Apple Intelligence need macOS 26.
  • The fn shortcuts only work on Apple keyboards.
  • The app interface is in German for now.

In collections

Topics

FAQ

What is blitztext-app?
Local-first macOS menu bar dictation app that transcribes speech with Parakeet or Whisper on your Mac and pastes it at the cursor. Blitztext is a dictation app for macOS. You hold a keyboard shortcut, speak, and release it; the speech is transcribed on your Mac and the text is inserted where your cursor is. Speech recognition uses Parakeet Ultra or OpenAI Whisper via WhisperKit, both running locally with Core ML on the Neural Engine. The app has no account or backend, and audio does not leave the device.
Is blitztext-app free to use? What license does it have?
blitztext-app is open source under the MIT license. Check the license terms before using it in commercial products.
Is blitztext-app still maintained?
Actively maintained. The last commit to blitztext-app was on Nov 6, 2026.
What can I use blitztext-app for?
Dictating emails, chat messages, and documents into any macOS app without sending audio to a cloud service. Turning rough spoken notes into cleaned-up written text. Transcribing in-person meetings on a Mac with speaker labels and a summary. Converting voice memos into text on a Mac.
How do I install blitztext-app?
Download the ZIP from the latest release page Move Blitztext.app to Applications, open it, and allow Microphone and Accessibility In the settings (Modelle), load a local speech model once, then hold fn + Shift, speak and let go
Where can I get blitztext-app?
The source code and installation instructions are on GitHub: https://github.com/cmagnussen/blitztext-app

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects