About this project

TypeNo is a minimal, privacy-focused voice-to-text tool for macOS. It captures your voice via a short Control key press, transcribes it locally using the coli engine (powered by Node.js and ffmpeg), and pastes the result into your active app — all within a second. No windows, settings, or cloud processing. Requires microphone and accessibility permissions on first launch. Supports drag-and-drop audio file transcription and lets you choose input devices. Built with simplicity in mind: one function, zero configuration. Currently in low-key maintenance mode — fully functional but not actively developed. Installation includes downloading the app and installing the coli CLI globally; model downloads occur on first use. Troubleshooting guides cover proxy issues and macOS accessibility bugs. Source available for building from scratch. Licensed under GPLv3.