This was Team Hydra's semester project for 11-411 Natural Language Processing at Carnegie Mellon University. This video succinctly explains what the project was about and what we did.
Somehow, we won 3rd place (of 17) in the question-asking aspect of the competition!
It is in the sort of state that you might expect after starting with a simple prototype, simmering ideas on the back burner (AKA procrastinating) for several months, then rewriting the whole thing in essentially a 3-day hackathon before the final deadline. I wouldn't recommend trying to run it, though the code should be sufficiently self-documenting to allow a determined user to figure out the Python module and Stanford NLP tool dependencies and command line args.
It is not (and probably will not ever be) maintained. The code is rather a mess. But I did learn a ton from working on this project, and might try my hand at a more mature implementation of this sort of application in the future.