paperless-ngx https://docs.paperless-ngx.com/ can do most of this as well… except for the automatic metadata retrieval. i’ve been really impressed with paperless over the past few days as i’ve set it up and imported gobs of PDFs from an abandoned Evernote install. The web interface is quite well done.

Interface really looks nice! We also develop a similar project but for indexing personal videos and images[0] based on CLIP. It is a Flask based application behind a caddy proxy. [0] https://github.com/ramanlabs-in/hachi

Interesting. However, if it can't handle creating citations for my research paper, I would need to stick with Zotero for managing and tracking the papers I care about. Ability to create and manage citations would be the killer app feature for many, I think.

Seems the same problem that TeX has been solving really well since several decades ago. Hyperlinks included

Installing now to check this out for my large collection of PDFs. Thank you for the open source library.

I've been using a self hosted instance of I, Librarian. If it does find the DOI number in the PDF it will query a database to get the metadata. It does a pretty good job so long as the correct DOI number is somewhere in the PDF. You can also batch import PDFs.

Ok had to look this up and thanks to the generic name it took a bit. Here [0] the link for everyone who comes after me. It's a simple(?) php based webservice.

[0] https://github.com/mkucej/i-librarian-free

Thank you for the tip. Will check it out now (child comment linked directly)

any update of usability here? I have a giant zotero library (and a folder of 100gb pdfs) that I would love to just switch to something like this...

I believe you are asking OP?

If you were asking me, this software is very minimal and not going to compete with zotero for an unknown time frame. What's your current need after using zotero? (Just curious)

---

That being said, I really appreciate these initiatives and applaud author of this library, didn't confirm but probably a one-person team.

No, I was asking you because you said you were going to implement it. I run zotero on two computers and sync them with unison -- but it just seems a bit fragile sometimes. I've corrupted the database once before and it was a hassle to restore. Sharing highlights from ipad back into the library is even more tricky: I sync the entire zotero folder (many gb, but only a portion of my ~200gb of pdfs) to rsync.net, and then sync specific folders through PDF viewer pro or whatever. Then my highlgihts sync to rsync.net, and then back to zotero.

The main benefit from zotero is automatic generation of metadata, renaming of pdfs, and syncing of highlights.

It just feels like a rube goldberg machine. I should actually just PAY 10 bucks a month for zotero unlimited itself.

Having a very rapid full text search functionality on top of all this is the dream. I haven't yet seen a super easy way to implement that beyond running elasticsearch, and that's a rabbithole I don't currently have the incentive to go down.

That chicken paper always gets a chuckle from me. In case you have never seen the video[0]. Unfortunately, it appears to have been recorded on a potato. Do not miss the audience question at the end. Beautiful academic satire.

[0] https://www.youtube.com/watch?v=yL_-1d9OSdk