This is very cool! I was just looking for such a project and ended up using quick-xml since it satisfied my use case (needed to stream certain matches from a big XML file), and I didn't see yours. I love that it is a drop in for libxml2, that must have been a challenging constraint.

I will seriously consider adopting it as long as it supports streaming matches (it seems like it does).

I’m quite interested in this. Where are you getting metadata and covers from? I found this pretty hard unless you pay for ISBNdb (OpenLibrary has bad covers, and Google Books API doesn’t give you high resolution ones I think).

Ha, openlibrary and google books for now. And yes, the covers are not great. Currently just pleased to get it working, will add better providers later.