Start recording
Use the Windows default shortcut, Ctrl+Alt+Space, or set your own binding.
For Windows x64
AudioBud turns a hotkey into private speech-to-text. Press, speak, and your words appear in the app or text field that already has focus.
The app is built for quick dictation into any focused text field, without switching to a recorder or uploading audio for processing.
Use the Windows default shortcut, Ctrl+Alt+Space, or set your own binding.
Silero VAD trims silence, then the selected local model turns speech into text.
AudioBud inserts the transcript into the active app by clipboard or direct typing.
AudioBud is a detached Handy fork with its own defaults, model choice, Windows focus, and frog-themed dark interface.
Download and switch between Parakeet, Whisper, Moonshine, SenseVoice, GigaAM, Canary, and Cohere engines from the Models screen.
Choose push-to-talk or toggle mode, microphone and output devices, audio feedback, overlay position, and paste behavior.
Format spoken numbers as digits ($25, 10%, 3.5), add custom words and deterministic word replacements, switch between formatted and raw output from the tray, and run optional LLM post-processing with your own provider key.
Keep recent transcriptions and recordings, set retention limits, retry entries, and clear data when you do not want it stored.
Parakeet V3 is the Windows default. Whisper remains available for broad language coverage and translation.
Fast multilingual transcription for 25 European languages.
Small, medium, turbo, and large variants through whisper.cpp.
Moonshine, SenseVoice, GigaAM, Canary, and Cohere cover speed or language niches.
Unsigned release: Windows SmartScreen warns on first launch because the installer is not signed yet. Choose More info, then Run anyway, or build from source.
Latest releaseBuild from source
The current release is validated on Windows x64. macOS and Linux code is inherited from Handy, but this fork has not validated those builds yet.
bun install
bun run tauri dev
bun run tauri build