What started as a broken Hardcover integration turned into a fairly substantial Bookshelf rebuild.
The Hardcover metadata backend was returning HTTP 500 errors, which meant Bookshelf could no longer retrieve or sync book metadata. I rebuilt the backend and connected it using my own Hardcover API access, and metadata syncing is now working again.
While fixing that, I also reorganized the setup into two separate Bookshelf instances:
- The original Bookshelf remains dedicated to ebooks.
- Ebook searching and management stay on the original instance.
- A second Bookshelf instance was created specifically for audiobooks on port
8788. - The audiobook instance was configured with its own root folder, audiobook-specific settings, SABnzbd, and Prowlarr.
- Existing authors and monitored books were copied into the audiobook instance.
- SABnzbd permissions were corrected so it can create and use the required download folders.
- The audiobook instance is now matching the existing audiobook files against its catalog.
The existing-file matching process is still running. Audiobook searches are intentionally disabled until that finishes, so Bookshelf does not start downloading duplicates while it is identifying and organizing files that are already present.
The simplest way to describe it is that I repaired the Hardcover “translator,” kept one bookshelf for ebooks, and built a second bookshelf specifically for audiobooks.
This became a much larger project than I expected, but both sides are now separated cleanly and the metadata integration is working again. 