Quant Voice

Quant Voice turns speech into text on the Mac. You press a hotkey, speak, and the recognized text is typed where the cursor is, in any app. It uses a Whisper model that runs locally through WhisperKit, supports Russian and English, and keeps English brand names in Latin script inside Russian sentences. A short press starts and stops recording, and holding the key records while it is down. It is for people who dictate in Russian or mixed-language text. The README says audio is kept in memory only for the phrase, no dictation history is stored, and nothing is sent to a server; the only network use is downloading a model you choose.

It requires macOS 14 or later; the README does not list chip requirements. It needs Microphone and Accessibility permissions, and dictation does not work until you download a speech model from the menu bar icon. The release is not notarized, so the first launch needs System Settings > Privacy & Security > Open Anyway, or right-click and Open. The app lives in the menu bar without a Dock icon, and the Quant name and logo are not licensed under the MIT code license.