iamcheyan/chordvoxmini

ChordVoxMini is a local, offline voice input tool forked from an open source project called ChordVox. Its main goal is to provide free speech to text that runs entirely on your device, with no cloud service, no paid subscription, and no collection of your audio or text data. The core workflow is simple: press a keyboard shortcut to start recording, speak, and the tool transcribes your speech and automatically pastes the result at wherever your cursor is, in any application. This means you do not have to switch windows or copy and paste manually, the text just shows up where you were typing. After transcription, a local AI model polishes the text to remove spoken language habits, filler words, and false starts, producing cleaner written output than a raw transcript would. There is also an experimental offline Chinese to Japanese translation feature, so you can speak in Chinese and have the translated Japanese text pasted directly, without going through a separate translation step or service. On Mac devices with Apple Silicon chips, the tool attempts to use CoreML hardware acceleration to speed up inference and reduce energy use, since running speech recognition and language models on a laptop can otherwise be slow and drain the battery quickly. The project is a personal fork maintained by a single developer and is described as lightly tested. Platform support has only been verified on macOS, compatibility with Windows and Linux has not been confirmed, so anyone outside macOS should expect rough edges or missing functionality. Setup requires Node.js and Python 3, installing dependencies with npm install, and running npm run dev to start the tool in development mode. The project is written in JavaScript and released under the MIT license, so it can be freely used, modified, and redistributed.