@aurynn It's not *exactly* federated, but something I've been working on experimentally is seekseek.org (more details: https://seekseek.org/technology)

The basic idea is to move away from generic keyword search and instead collaboratively maintain search engines for specific purposes, that pull from known-good sources.

"Known-good sources" can be pretty widely defined there, and vary from search domain to search domain. It'll plausibly involve some amount of collaborative manual index maintenance at some point (along the lines of Wikidata, OSM, or Open Food Facts).

This doesn't exactly solve the *same* problem as generic text search engines, but it should be strongly resistant to SEO and similar bullshit, and make for much more accurate searches within supported 'search domains'. And to be honest, I think there are way more interesting possibilities *outside* of the "let's just clone what Google does" space anyway.

Federation isn't currently a part of it (and I haven't yet seen a spam-resistant model proposed for that either!), but being locally replicable and divided into domain-specific search systems sort of naturally encourages for different groups to run different parts of the ecosystem.

It's still pretty early days, and I'd love for more people to get involved! But the basic concept seems to work so far.