Use SearXNG in Firefox
Nov 29, 2024 by @turgon@dosgame.club(You can skip my nonsense and head straight for the instructions.)
Yesterday I posted to Mastodon that I was going to look into running my own instance of SearXNG, but ended up deciding to take the doc's advice and use an existing instance for a while first.
I remember when using Google was exciting. Its predecessors had built monstrous, cumbersome taxonomies and searching sucked. It felt like they had approached true information scale and just couldn't hang with it. Trying to find something very specific usually took a lot of time and effort. (Sidebar: I spent hours scouring Yahoo in the 90s for music and found Mogwai and other bands this way, and it felt amazing.)
Then Google stepped in with a better algo and simple single line of text search input and took over. This also felt amazing because we could find more stuff with a lot less effort. I'also nostalgic for recording songs from the radio to make mixtapes, but that doesn't mean I want to spend my time recording the radio today.
Some people think the web should go back to blogs, webrings, and deeply thoughtful curated collections. I couldn't agree more. But I see a lot less conversation about how to find that content. I was pretty excited to learn about SearX some time ago and wanted to set some time aside to look at it more closely.
In the meantime I switched my default search engine on my phone and personal computers to Ecosia, which feels like a pretty mid choice after a year, but at least they weren't jamming AI in there (which has since changed). Ecosia seems to be outright trying to become Google, but where Google once said "don't be evil," Ecosia says "we'll plant a lot of trees."
What I really want from a search engine is:
- to find the best things, of course! and quickly!
- not to track me and what I search for
- not to show me ads
- seriously just do search and that's it
SearXNG itself is not a complete solution. It's really just anonymizing and aggregating my searches across many different search engines. Information about me and my searches are kept private from the search engines, but not to the SearXNG instance admins. In the results, I can click through to a page, but I can also view the SearXNG instance's cached view of the page, or proxy the request to the page through the instance. So there is even some control over the personal/tracking information presented to the pages I view from the results.
Alright, on to the first step in setting up your browser: finding a SearXNG instance.
Start with the list of public instances. I recommend filtering for your country and for A+ TLS and CSP fields, which you can do by entering values in those inputs just below the column headers. Then you can click into the instance and try searching for stuff with it, or click the link to "Engine Stats" below the search form. This will list all of that instance's sources for search.
Once you've got an instance picked out, follow these instructions to set it as your browser's default search engine. In desktop Firefox, while visitng the SearxNG instance, right click in the URL bar and a little menu will popup, including an item to add the instance. Once you've added it, open up Firefox's Settings, find Search, and set it as your default search engine.
The process was similar with my phone (android), but I had to paste in the SearXNG instance's URL and append ?q=%s
to it, which is weird because how else would it even work.
And that's it!
SearXNG instances cost money to run and I think it'd be great to help support those instance costs just like I support my Mastodon instance. Not all the instances listed give you the ability to donate, but some do. Ononoki for example has a donate link at the top right of the page (a little heart icon on mobile), which lists a few ways you can chip in. I paid them $26, their "medium" level patronage for six months. Hopefully by then I'll run my own instance. Either way, I am perfectly happy and able to pay for using search as a service rather than get ads up in my face all the time.
Finally, when I said SearXNG wasn't a complete solution, I meant that there is still a need for more search engines that specialize in search for particular kinds of content. Curation of content by experts or at least fans who genuinely love it should be informative to search. That is, if that collection had a custom search engine I would expect it to perform better than a general purpose search engine like Google, and bringing that engine into my search aggregator should be a good thing. Search should be federated!