A native macOS app for speech-to-text transcription. Record audio with a global keyboard shortcut, transcribe it using OpenAI Whisper or Mistral Voxtral, and insert text into any active app. Features a minimal menu bar UI, customizable settings, and transcription history.
Because I dindn't find any existing app that would support Mistral's Voxtral 🤷♂️
-
Global Keyboard Shortcut: Toggle audio recording with a customizable hotkey .
-
Speech-to-Text Transcription: Supports OpenAI Whisper and Mistral Voxtral APIs.
-
Text Insertion: Automatically inserts transcribed text at the cursor position using Accessibility APIs.
-
Menu Bar Interface: Displays recording status and provides quick access to settings and history.
-
Settings: Configure API keys, transcription service, insertion method, and keyboard shortcut.
-
Transcription History: View, copy, or delete past transcriptions with search functionality.
-
Onboarding: Guides users through initial setup, including microphone and accessibility permissions.
- https://superwhisper.com/
- https://betterdictation.com/
- https://spokenly.app/
- https://tryvoiceink.com/
- Allow model selection (not just provider). Useful for gpt-4o-transcribe or gpt-4o-mini-transcribe
- Add ElevenLab's Scribe
- Allow Custom vocabulary or custom prompt for elligible models
- Queue service for async processing of transcription requests when offline
This app is 90% vibe coded. I never coded in Swift.
- Uses Apple's Unified Logging with
os.Loggervia a small helper inVoxMac/Log.swift. - Categories:
app,audio,transcription,history,database,permissions,notifications,ui,insertion,network,keychain. - View logs in Console.app: search for the app’s bundle identifier and filter by category/level.
- Debug messages are verbose and best viewed during development; info/notice/error/fault are useful in production builds.