What it is
Handy turns speech into text on your own computer and pastes it wherever you are typing. Hold or tap a shortcut, speak, and release: it filters silence, transcribes with NVIDIA's Parakeet V3, which runs well on an ordinary processor, a Whisper model (faster with a graphics card) or Cohere Transcribe, its most accurate English model for machines with a capable graphics card, and inserts the result. Transcription never leaves the computer. It runs on Windows, macOS and Linux, and the code is MIT-licensed on GitHub.
Handy is free and by default pastes a plain transcript. An experimental post-processing setting can send that transcript to a language model to clean it up, using OpenAI, Anthropic, Groq or another provider with your own key, Apple Intelligence on recent Macs, or a local model; only a cloud provider takes text off the computer. It doesn't adapt to the app you are in the way the paid dictation apps do, and there is no phone version.
Best for
- Free dictation on Windows or Linux with nothing uploaded
- People who want an open-source tool they can inspect or extend
- Pressing one shortcut to speak into any app
