voiceink

VoiceInk is a native macOS application that transcribes speech to text using local AI models, designed as an open-source alternative to Superwhisper and Wispr Flow. It requires macOS 15.0 or later and is licensed under GPL v3. Core capabilities: - Local AI transcription powered by whisper.cpp, SenseVoice Small (multilingual model by FunAudioLLM/Alibaba), and Parakeet models via FluidAudio, with offline processing that keeps data on-device. - Modes: intelligent app detection that applies pre-configured settings based on the active app or URL. - Context awareness: adapts transcription based on screen content. - Global shortcuts: configurable keyboard or mouse shortcuts for recording and push-to-talk. - Personal dictionary: custom words, industry terms, and smart text replacements to improve recognition of unique terminology. - Smart modes: switch between AI-powered modes optimized for different writing styles. - AI assistant: built-in conversational assistant mode. Installation options include downloading from the project website with a free trial, installing via Homebrew (brew install --cask voiceink), or building from source using instructions in BUILDING.md. Purchasing a license provides automatic updates, priority support, and helps fund development. The project is not currently accepting pull requests, but contributions via bug reports, feature suggestions, and documentation improvements are welcome. Key dependencies include Sparkle for updates, KeyboardShortcuts, LaunchAtLogin, and SelectedTextKit.