Readarr from the -Arr team is slowly shaping up and other alternatives such as LazyLibrarian have been available for a while.
While I think it would make most sense to focus on the implementation of Readarr, it is still far from finished.
It would be nice to first implement the request type without necessarily having the backends implemented yet, and use GoodReads (or whichever "database" there is for books really) for the references.
This would already allow me to "take" requests and fulfill them manually until a backend would be implemented
(p.s. i swear I made a request for this but I cannot find it using search, not sure where it went. Please delete it if you can find the original)