Should fragments include metadata?

Right now I'm just pulling down a single document, as text or PDF, and storing that as context. But we have other user-controlled metadata we could include -- title, description, etc -- that comes back in the API call. Should we include that, and how?
