VoxKey

VoxKey is a free, open-source dictation app for people who want to speak into Mac text fields without sending audio to a cloud transcription service. It requires Apple Silicon and macOS 26 or later. Speech recognition runs locally and works offline after you download a model.

The native menu bar app inserts dictated text where you're writing, with a keyboard trigger for starting and stopping speech capture. Built with Swift and AppKit, it runs its models inside the app, so you don't need a separate local server or runtime. Safety Net lets you recover text if the app can't confirm delivery to the destination editor.

WhisperKit powers speech recognition. Distil-Whisper v3 models handle English, while multilingual Whisper v3 Turbo models support Spanish and other languages. You can keep several models locally and switch between them.

Optional English cleanup removes fillers and repetitions and corrects grammar using Superwhisper's S1-mini and GECToR. It's off by default. Both models process text on your Mac, but the grammar model permits noncommercial use only, and corrections can change your words.

VoxKey doesn't keep recordings or a dictation history. Audio, transcripts, preferences and downloaded models stay on your Mac; model downloads and update checks need internet access. After text reaches another app or the clipboard, that destination controls how it handles the text. VoxKey's source code uses the MIT license, while model assets have separate terms.