Bookshelf – Self-hosted eBook library that runs on object storage

Let's be real: the self-hosted ecosystem is drowning in half-baked "media server" projects that all want to be Plex for books, or Plex for podcasts, or Plex for your grandma's recipe cards. Bookshelf cuts through the noise by making a bet most of these projects are too cowardly to make: that object storage is the only backend you actually need. No Postgres, no Redis, no SQLite files that get corrupted when you look at them wrong. Just a bucket, some keys, and a web UI. That's it. The architectural discipline here matters more than the feature set. By tying the library to object storage, you get the holy trinity of self-hosting: cheap storage, infinite scalability (if you want it), and portability across any S3-compatible provider. If you ever want to jump from MinIO to Backblaze to a cheap VPS disk, you're not migrating a database—you're just pointing a config at a different URL. That's the kind of freedom the "own your data" crowd should be clamoring for, but instead they're still fighting with Calibre's monolithic library directory and its SQLite tables that lock you in. Of course, the project is early. There's no word on metadata scraping, format conversion, or the million little features that make a book library actually usable. And object storage has its own weirdness—latency on cold reads, no filesystem semantics, and the fact that some people don't want to learn a whole new mental model just to store EPUBs. That's the trade-off you make for simplicity. But here's the thing: every few years, a project comes along that reminds us the simple path is the one we should've been walking the whole time. Bookshelf is that reminder for the book-hoarding crowd. It's not trying to be the most powerful thing in the room; it's trying to be the easiest thing to keep alive for the next decade. And in the self-hosted world, that's a feature that beats a dozen extra bells and whistles. Check the repo yourself: https://github.com/murerkinn/bookshelf