This already exist: "searching over the titles, text and URLs of sites you've visited before" -- https://www.browserparrot.com/
And you think that hasn't been made yet? :-D
Show HN: DeepHN - https://news.ycombinator.com/item?id=26791582
Of note, I was able to find that item (which I recalled existed but not by what name) with hn.algolia, but I was not able to find it with the OP search engine, or with the DeepHN search itself. So in my book, Algolia is winning.
But projects like this are super fun and educational to build so props to OP.
OP said they only indexed the past 6 months.
Perhaps someone will be able to help me find a website I’ve been trying to Google for with no luck: the main focus of the website is a music player with 3D pipes and arrows that the camera would follow along, live rendering the entire scene while music was playing, but it was also a personal blog if you scrolled down. Domain was relatively short and I think it started with K.
Currently it's indexing the content of stories and comments.
Are you suggesting also to index the content of the main story link (which is outside HN)?
That is what they're suggesting, yes.
maybe add metadata to the index, but not the full page content, else you're just gonna build another google
How so? There's no PageRank in sight here, very little spam to worry about or contend with.
It would be remarkable and interesting to have a super deep search capability that indexes all first-order links on this site.