I think this is way overdue 
Currently, the only way we can give instructions to the user is through the fixed list of prompts or graphics pre-recording. Sometimes the prompts don’t really have anything that is suitable for our filter and when we do graphic instructions with English, we are excluding regions in the world that do not use English as their main language.
The text fields can allow us to have localized instructions (hopefully) and also allow us to create more optimized effects without the need to export a graphic for every single letter or number in some cases.
There is also so much potential we can explore with text fields to make our effects more interesting!