i am so confused when i search for this info, because the limitation all lead to the google service "Speech To Text" api , and i think it is also the same service behind the desktop chrome speech recognition