What is voco?

Voco is a native macOS voice input app built with SwiftUI that turns your speech into text inside whatever app you are already using. It lives quietly in the menu bar rather than as a full window, and the main settings window only opens when you need it. To use it, you press a hotkey and start speaking. A small heads up display at the top of the screen shows a live, partial transcript as you talk, so you can see the words appear in real time. Once the transcription is done, the finished text is inserted directly into whatever app currently has focus, whether that is a note, an email, or a chat window. Voco supports two ways of turning speech into text. One option is Volcengine, a cloud based transcription service, using either its newer API key system or an older app ID and access token method, with credentials stored securely in the macOS Keychain. The other option is a local model that runs entirely on your Mac using the sherpa onnx streaming runtime, so no audio needs to leave your device. The recommended local model handles both Chinese and English and downloads automatically the first time you use it. The app also includes a statistics page that tracks things like how much you have dictated, which apps you use it in most, and when you tend to use it, along with a basic skill for cleaning filler words like um out of transcripts. Voco needs microphone access to record audio and accessibility permissions to insert text and listen for the hotkey system wide. It requires macOS 14 or later on either Apple Silicon or Intel hardware, is written in Swift, and is released under the MIT license.