I setup the desktop with Japanese locale and simply followed the steps in https://fedoraproject.org/wiki/Changes/ibus-speech-to-text_pywhispercpp 1. install ibus-speech-to-text 2. ibus restart to recognize the installed engine 3. add Speech To Text to the input sources 4. Select Speech To Text with Super+space to activate and to access the setup from ibus menu After those steps, the STT setup window is open. That's fine because I have no models downloaded. I see a warning/information like "There is no available model for the current locale" and "Download Model" button, but anyway. I changed the language to en_US and download models and select it, and close the setup window. When I deactivate STT with Super+space and activate again, the setup window appears again. Even though I have one models for selected "locale" there. They still complain that "There is no available model for the current locale". Reproducible: Always Additional Information: Apparently it get back to the normal after rebooting. Though "There is no available model for the current locale" message is still there. Strange.