A community project is a great idea.
Both A and B are good projects. I wonder if there is any interest in a third option, C.
Over the years on this forum I have seen people interested in Speech Recognition. Both for recognizing Commands for an app, and for Continuous Speech Recognition to translate speech to text continuously as it's spoken.
It's amazing how complex either of these problems is once you get into the details, and I think both have numerous real-world applications and advantages for B4A developers.
An effective implementation of Command recognition or Continuous Speech Recognition is really very similar, requiring the same technologies.
This could become a very long post …
Is there any interest in a community project tackling something like this?
Barry.