TalkToPaste

TalkToPaste is a voice dictation utility for macOS and Windows that runs speech recognition on your computer. The official site links to the VoiceToClipboard repository, which documents the local implementation and installation scripts. Transcription works offline after the speech model has been downloaded.

Hold Left Control and Left Option on macOS, or Left Ctrl and Left Alt on Windows, while speaking. Releasing either key starts transcription and copies the result to the clipboard. You then paste the text into your email, document, chat or code editor. A menu bar or system tray icon provides settings, and a floating indicator shows recording and processing status.

Speech recognition uses Whisper through whisper.cpp. Models range from Tiny to Large V3, with Base selected by default. Metal acceleration is available on macOS and CUDA on Windows when supported.

The documented installation runs from source with Python and platform build tools rather than a verified packaged installer. macOS requires version 12 or later and Xcode Command Line Tools, with Apple Silicon recommended. Windows requires Git, CMake and Visual Studio Build Tools. The README describes the project as MIT licensed, though the repository metadata does not identify a licence file. All documented transcription processing stays on the device.