Okay folks, a word on the vibe-coding thing, since I can see it stirred up a lot and clearly rubs some of you the wrong way. Let me just drop a few numbers so you have an idea of what this actually is.

I started this project in February. I work on it mostly in the evenings, after my day job. Over ~4 months I estimate I’ve put around 100 hours into it. I use it every single day, and I’ve tested and optimized it quite a bit. So no, this isn’t some thing I threw together in 2 hours with zero investment.

Yes, I built it with AI in the loop. But without it, I’d never have had the time to make something this “polished” on the side. And honestly, why would I deny myself that on a personal project? When a carpenter builds a piece of furniture, I don’t hold it against him for using a power drill instead of a hand brace.

That’s it, just a small rant. I won’t engage with the AI criticism beyond this. Back to talking about the actual software for anyone interested. Cheers.

Funny timing, I’ve been playing with Jellyfin lately too and I ended up forking a little project to bridge Plex into Jellyfin: https://github.com/ndieschburg/xtream-to-strm-web

It syncs your Plex.tv libraries (even shared servers you don’t own) into .strm and .nfo files, so Jellyfin just picks them up as a normal library and you can watch everything from any Jellyfin client. There’s a built-in proxy that hides the Plex tokens from the STRM files, and for clients that don’t follow HTTP redirects (like Findroid) there’s an HLS proxy mode so it still plays fine.

Still a bit rough but it does the job if you want to keep a single Jellyfin frontend and still reach the stuff that only lives on Plex.

Thanks for the suggestion. I know Audiobookshelf, but it’s audiobook-first and the EPUB side is basic, it doesn’t do the KOReader ↔ web position sync I built this for. And no worries about the AI part, I was upfront in the post on purpose. You don’t have to use it or like it. I built it for myself, it works for me every day, and I shared it in case it’s useful to someone else. That’s all.

Touché 😄 Honestly that’s the one feature I can’t compete with: a book has infinite battery and zero downtime. Best I can do is offline caching.

Not a problem for me, because my sync is manual, not automatic. I assigned it to a gesture (tap the top-right corner): wifi on, sync, done. Waking the device doesn’t trigger anything, so I never get that “jump to latest position?” dialog.

I did it this way mainly for battery, I don’t want wifi turning on every time the Kobo wakes up. And I push my position at the end of my reading, so the server is always up to date and there’s nothing stale to pull next time.

I did think about auto-syncing when you open a book, but haven’t done it yet, partly for the reasons you mention. For now the manual gesture works well for me.

Of course sir, here are the keys: API_KEY=aGVjayBvZmYgbWF0ZQ== (decode it if you dare)

Same energy as “I don’t need GPS, I have the stars” 😄 But fair enough, no problem if a real book works for you! The hard part of my project is not the library, it’s syncing the reading position between my Kobo and my phone. I put a lot of work and testing into that part. It’s been my daily reader for 3 months. Code is here if you want to look: https://github.com/ndieschburg/varbook