@andrei_jay hmm you mean internally? I want the textual representation of both sides but you are right that a voice to text or text to voice app does need this sort of information so it must have a layer somewhere