Actually isn’t it something that’s supported by operating systems? i.e. some utility that runs, records the speech, and sends it to whatever text input is currently active. I think it would make more sense than implementing this on every text editor, but not sure if such an utility exists.
