Back in the day the best way to find cool sites when you were on a cool site was to click next in the webring. In this age of ailing search engines and confidently incorrect AI, it is time for the webring to make a comeback.
This person has given his the code to get started: Webring
So, classic webring navigation consists of arrows to the next and previous ring instances, as well as a link to the ring index. By their nature, webrings are manageable-sized communities by nature. I don’t see how that can be improved upon by a search function?
well the central site of the web ring could be searched for any particular page that’s part of the ring, and that search could be surfaced on any page that’s part of the ring.
The full set of pages could be decentralised and cached across all members for robustness, and even include each page’s own description and recommendations for every other page if they like.
And then, of course … rings of webrings with as many levels of aggregation as people are interested in maintaining, again with decentralised caches of pages, their links and descriptions (all human curated of course) that can all be searched whenever a member page or aggregating page opts into it.
Tech capabilities have advanced since the 90s enough now that basic text search in a web page over a small data set is not hard or too much to ask.
And nested rings of rings of rings are scalable because at each level the data will just be links (and descriptions or names if available) while it would be on the user to navigate the various layers however they wish until they find something they’re interested in.