Typester enables speech-to-text dictation directly into any macOS application through menu bar integration. The app converts spoken words into typed text using configurable speech recognition providers without requiring cloud intermediaries or subscription fees through the app itself.
Users choose between Soniox or Deepgram for speech recognition services. Soniox provides support for over 60 languages with custom dictionary functionality for domain-specific terminology. Deepgram offers automatic language detection for multilingual workflows. Users provide their own API keys and pay consumption costs directly to their chosen provider.
Two activation modes accommodate different dictation preferences. Press-to-speak mode activates transcription while holding a configurable hotkey. Toggle mode enables continuous dictation with single hotkey activation and deactivation. Real-time transcription uses streaming APIs for low-latency processing.
Microphone selection allows users to choose preferred input devices from the menu bar. Auto-paste functionality automatically inserts transcribed text into active applications using simulated keyboard input. API keys store securely in macOS Keychain. Launch at login enables automatic startup for continuous availability.
Built entirely in Swift with a native SwiftUI interface. The app uses WebSocket streaming clients for both providers and AVAudioEngine for microphone capture. Global hotkey management enables system-wide activation regardless of active application.
System requirements: macOS 13 or later. Requires valid API key from Soniox or Deepgram. Swift 5.9+ required for building from source.
Permissions: Microphone access required for voice capture. Accessibility permission required to simulate keyboard paste (Cmd+V) into other applications.
Pricing: Free and open-source under MIT License. Users pay only for API consumption directly to Soniox or Deepgram based on usage. No middleman fees or app subscriptions.
Limitations: API credentials required before use - the app does not function without external service configuration. Custom dictionary support limited to Soniox provider. Auto-paste requires Accessibility permission, which may raise security considerations for some users.
Alternatives: Users seeking built-in solutions may use macOS native dictation (limited language support). For commercial speech-to-text with app-managed subscriptions, Dragon Dictate offers comprehensive features. Whisper-based local solutions like MacWhisper provide offline processing without API costs.
Suitable for professionals, developers, writers, and accessibility users seeking efficient voice-to-text workflow across macOS applications. Best for users comfortable managing API credentials and preferring direct provider relationships over app-mediated services.
