It suits users and developers who need offline voice input without sending recordings to a cloud transcription service. Core capabilities include English speech-to-text, A 163 MB base download, Local Mac inference, and Clean variant removes fillers.
Public code or model weights are linked from the project, with setup instructions for running or adapting the system. Additional characteristics include Plus variant improves verbatim accuracy, Pro full-precision variant, Dictation in Underdog Speak, and Meeting-note transcription.
