Calibre Web Automated is what I use. Works well.

I’ve been using Grimmory(previously Booklore) and have been very happy with it. It integrates well with Hardcover and Kobo ereaders (with some minor software modifications).

I don’t currently have an automated solution for downloading ebooks.

I’ve been pretty happy with Kavita as the server. You can read the books directly from the web app but if you think you might end up in an area with no signal, you can download the book from Kavita, and use a separate eReader app to read the file locally.

I’ve got BookOrbit and Audiobookshelf both going. They both can be hosted securely (locked down compose file with read-only, non-root user, etc.) and use Postgres as their DBs, both key features.

I added in BookOrbit to try since it has kobo sync and koreader sync that Audiobookshelf lacks.

I moved books from Audiobookshelf to get BookOrbit going and there was a learning curve to get the UI to do it optimally for me, but I eventually got it to work for me. BookOrbit has the ability to write metadata to the files themselves, which most things lack. Very nice for portability.

There’s a folder BookOrbit imports from and you can set it to populate metadata automatically - seems strongly built for an automated library system.

Both have been very stable. I’d say BookOrbit is the better one - and it supports audiobooks too. Audiobookshelf handles multiple libraries (like books and comics) in a clunky way (have to switch between them like they’re completely different silo’d libraries - much like how Calibre handles them). BookOrbit has them separated but easy to see they exist and you can mix and match them in a collection or something. Better way to handle it.

I use the desktop application Calibre to convert books as needed, but BookOrbit will automatically generate kobo epubs from epubs when syncing so I need not worry about kepub prep.

Lastly, I chose BookOrbit to try over others because Grimmory needs a ton of RAM, Kavita had features behind a paywall, some other one is comic-focused, and the Calibre web iterations give off the vibe of a lot of tapes the inside to make them work; I had big doubts Calibre Web Auto would be able to be run non-root and read-only. Chose Audiobookshelf originally because of the Calibre mess and other options didn’t exist or were much less established.

Edit: lore drop: BookOrbit is a feature copy of Booklore but written not in Java (I think JS), and Grimmory is a community fork of Booklore after its creator fell into AI psychosis.

I just use Audiobookshelf because i dont want to have two apps for books 😀

I’m very happy with Kavita running in Docker. I’ve used it through the Web UI and KOReader on both phone and Kobo. All work great. I manually download books and fix the metadata on them though so YMMV.

Readarr was orphaned, so most people switched to using LazyLibrarian. LL is kind of difficult to work with and you’ll typically see a lot of failures in the logs for various things. Instead, you could use Shelfmark which fulfills the same general purpose but is more straightforward.

For browsing/reading, most people use Calibre-Web. You can write a simple bash script to do periodic imports via calibredb. Just point the output directory of Shelfmark to the input directory for that command.

EDIT: I’m not sure if this exists in Headscale, but Tailscale also has Services you can setup so that something like machine.some-domain.ts.net:8083 can be mapped as books.some-domain.ts.net. For Kobo Sync or OPDS devices you can just download to local, but if you are just using an Android device, you can stream the book to the device via browser which is very nice.

Could look at mirroring Project Gutenberg.