[dekarrin]

Yeah, it's gotten so bad that I'm starting to seriously contemplate rolling my own web crawler and indexer.

The logical side of me says that's going to be way too much work to manage on my own. But it's hard to shut down the nagging feeling that pops up every time Google chooses to disregard a "+" or double-quotes around a word. That feeling of "would it really be that hard to toss a bunch of site data into an ELK stack and implement a primitive PageRank-ish score on sites? Hmmmm..."

Maybe just a YouTube indexer (whose search has also gone to complete garbage these last few years). Though, that would have the disadvantage of not being justifiable via eventually helping to make me more productive.

[toomuchtodo]

Crawl the internet archive Wayback machine slowly. Once you have built your index, you can fill your corpus gaps with public crawls. Patron services can advise to be helpful without causing stress on their infra.

[jeffrallen]

I use this and it works great: https://notrashsearch.github.io/about.html