I see searx-admin was done and being packaged in 2020. Is this available for Debian 11-based builds and, if so, how do we install/access this administration interface?
I have a look at the settings yaml file and the old threads on this here are right. That is a mess to read and difficult to understand how to effect the changes desired. I’d love to be able to configure my searx instance for all users, vs using the web link or saving cookies on every device for every user.
I was also interested in having more access to the Searx features. Specifically, I wanted to be able to get more than one result per engine to show up; I feel like the way it is set up OOTB is not that useful.
I attempted to customize my Searx instance using the documentation available here, but it ended up breaking Searx for some reason.
Actually, even replacing the FreedomBox
settings.yml with the default 'settings.yml` posted on the GitHub page (here) breaks Searx. So there is something about the way it is being executed on FreedomBox that deviates from the “standard” instance.
All that to say: I would also appreciate if there were an admin package to help configure Searx!
@BluishHumility I think it might be because Searx in FreedomBox, being a Debian package, is very much behind the “standard” upstream version.
That does make sense. I did get the sense the FBX instance is customized a little bit under the hood, but it could be the documentation doesn’t apply to the version of the application we are running as you mentioned.
What started me poking around with the
settings.yml file in the first place was I had used a public Whoogle instance a few times and I really liked it–the results were just better than what I was getting from the default Searx configuration. I was curious if I could get my Searx to give me similar results by leaning a little more heavily on the Google engine.
I started off by making Google the only engine, but then for some reason a query only gave me a single hit on each search. After digging around trying to figure out how to broaden that to a more useful number of hits I kind of hit a wall–the GUI interface is very limited option-wise, and there is no admin panel as OP mentioned. I wound up trying to achieve a better configuration in the
settings.yml file, but it just didn’t tolerate being modified in ways the documentation suggests it should be able to.
I ended up setting up a Whoogle instance in a Docker container on a different Debian box and it’s working out great, but I do think it would be good for the community if we could get the Searx application more deeply configurable–without getting too in the weeds with documentation that may or may not be applicable to the version we are running.
I started off by making Google the only engine, but then for some reason a query only gave me a single hit on each search.
It’s worth setting the language under “Advanced settings” on the right side. In my experience, it improves both the quality and the quantity of the results.
I ended up setting up a Whoogle
I like Whoogle, but when I was hosting it, the EU cookie terror broke it too often, so I gave up on it.
Wow, that looks super annoying! But how did it break your Whoogle? And also how does Searx get around it?
Here is an example: [BUG/FEATURE] Bypass EU cookie consent · Issue #243 · benbusby/whoogle-search · GitHub
I don’t know what Searx does differently but it doesn’t break fortunately.