I also host my own MediaWiki instance, complete with VisualEditor and a mobile skin. Out of curiosity, why are you not loading documents and images into your wiki? For images, I don't use it as an image repository, just for images of things specific to personal knowledge.
Well in general I want documents to be available locally. If I were to upload them to the wiki, I'd had to 1) download them from somewhere, 2) upload them to the wiki, 3) when I need them, download them again for the wiki. In contrast, if I use upspin, I just download it in a directory served by upspin, so it's just step #1 above, and I always have them available.
I do upload pdfs including bills and invoices (not paychecks since I don't have a use case for that). My wiki is private so it's only available to me. Plus I have extensions that allow me to render the pdfs nicely and index all the text in pdfs for easier searching.
I have both a public and a private wiki, but most of my PDFs are 5000+ page instruction set architecture monsters, so I much rather prefer to view them offline in some decent viewer instead of viewing them in the browser.
I also have passports scans and other sensitive stuff. I don't trust that I can secure mediawiki well enough to put those in clear on the internet (in upspin they are encrypted with a local key and the storage is my own server, mediawiki is in the public cloud).
I've never actually heard of upspin before this thread. Their homepage doesn't give a clear view of what it actually does, but I'm looking over https://upspin.io/doc/overview.md now
What do you use for indexing/searching pdfs? I didn't think about that even being a feature for something like mediawiki, but I guess that would make sense.
One of the big features I like in Evernote is the ability to search pdfs and images. I have a lot of technical manuals that I save, along with receipts/etc that are easier to search for.