Can "text chat as the input and voice as the output" be recorded as a feature request? It would be fantastically useful!