I’m currently working on re-evaluating our search engine selection (reading privacy policies and all that good stuff), to see what to keep, remove, maybe add. I figured I might use some input from lemmy.
- what do you use out of the ones we include? is anyone actually using search engines like qwant and metager?
- do you add any search engine to librewolf?
if you’re curious bout my notes on this -> https://gitlab.com/librewolf-community/settings/-/issues/111
Currently, Im using duckduckgo. I love the bang feature. Im curious about decentralized engines like searx.
I personally use SearX as my primary search engine (one local instance, the default searx.be instance as a fallback). Sometimes, I use one of the following search engines as well, but generally, there is no need for them in my case (for some of them, I retrieve their results with my SearX search, anyway). I rarely get to the point where SearX cannot provide the required and satisfactory results and I have to search with some specific search engine. The search engines I currently have set up in all of my browsers on both desktops and mobile phones are:
- SearX,
- DuckDuckGo,
- StartPage,
- Qwant,
- Metager and
- some specific search shortcuts for a few websites (Wikipedia, Lemmy, …).
I have one Whoogle instance set up too, just for testing purposes, though. And I love the idea of per website aggregated results provided by Gigablast, but Gigablast is way too slow to use regularly. Interesting idea, though.
Whoogle seems to get blocked very often for me, both selfhosted and public. I’m actually thinking whether we should be remove it and put back Startpage because of that.
other than that, google-only results is not my cup of tea, so just like you I mostly have it for testing.
Qwant
how good is it? it’s another one I was looking into removing.
Whoogle seems to get blocked very often for me, both selfhosted and public.
It seems to be the same for me. I used to get timeouts or other errors when testing it. And as far as I can say, it lacks a lot of features I find necessary nowadays (advanced filtering, options etc.). And yeah, if possible, I try to stay away from Google results, or at least mix them with others. Thinking about it, I would be up to remove Whoogle in favour of Startpage.
I actually really liked Brave Search for their results mixed with their own indexing. I have no idea how good Brave Search is privacy-wise (it will not be such a win, I can think), but the results are definitely nice in my opinion.
Qwant
Furthermore, I find myself quite enjoying Qwant in general. The results are good, and the functionality is ample. What I really hate about Qwant is that it is not OSS. Qwant being a European (French) company with respectable reviews (there is always something, of course, but I do not recall that much of a controversy about Qwant) gives it some credibility in my eyes (although, credibility might be a bit too strong of a term). I trust them to some degree, though, for sure. And I used to search Qwant exclusively when I could not find something (especially for national searches) at SearX quickly enough. It worked surprisingly well for me.
The same goes for Startpage (European roots from the Netherlands), for example, but to be honest, If I have to choose between Qwant and Startpage, I would probably go for Qwant. But I did not test Startpage on its own that much, to be honest. It might just be much better than Qwant for all I know. I think I might give it a try for some time…
From my perspective, I would keep Qwant and treat it the same as Startpage, for example. I do not see that much of a difference between these two. I would highly appreciate if anyone here knows more about Startpage or Qwant and point me to some privacy-oriented facts considering these search engines. It would be great to learn more. All I know is just what I have been able to pick up here and there. I lack deeper insight into all of this, sadly.
for evaluation of privacy policy and practices of each search engine you can see the link in OP, I’m done evaluating all of them. Startpage and Brave search actually came out cleaner than Qwant, but overall all of the evaluated ones are very much above average.
I think Startpage main advantage is that it’s a Google alternative, and atm it seems like a more viable one than Whoogle, which we tried to push instead. Brave and Searx are cool, I don’t see much added value that justifies including them by default (if one likes the results just add it manually), so if we were to trim some stuff I would consider them honestly, but I’m not sure either.
To add to the Gigablast mention, they teamed up with freenode to create private.sh search engine… Well, I do not know whether they mean the former freenode (nowadays Libera.Chat), or the new one (the freenode after that debacle with the takeover of freenode). If the latter is the case, well, I would say stay miles away from it. And considering I have not seen any mentions of it anywhere, I would be marginally cautious. It might be all all right, but who knows.
From private.sh How it works site:
Built to be audited
Rather than relying solely on closed technologies, we’ve made a genuine effort to publish our code and to use well known and thoroughly tested open solutions to provide a truly private search experience.
Our website and extension is built with open source technologies and can easily be audited and improved through our open source program (coming soon!).
Well, I do not know, but I have not been able to find the source code, so as far I can tell so far, I cannot believe them.
I will look into this, thanks!
metager.org atm
I use searx disroot instance. There is private.sh which is interesting, there is ecosia.org which I like though they ought to be compliant with LibreJS. Ecosia is essentially bing and tracks ip too but it functions well with tor so I sometimes use it.
Ecosia. I am willing to sacrifice some privacy to support their cause. For sensitive searches I might not want associated with my user profile I usually use duckduckgo through Tor.
deleted by creator
searx, but I do not self host. Currently lusing searx.be most of the time…
I myself just use the default one (DDG), I’ll watch this thread if someone posts compelling reason to switch.
I’ve really grown to like ddg. I used exclusively searx in the past, but now I mix.
in the end the final decision was:
- keep ddg, ddg lite, searxng, metager, wikipedia.
- remove whoogle, brave, qwant.
- add startpage.
basically, reduce the number of engines by keeping only the most popular ones (nothing against the others, but users can add them manually), and replace whoogle with startpage for google results.
by popular ones I mean the one that were suggested the most in this discussion, so thanks everyone!
I am using a searx instance.
do you use the default librewolf instance (searx.be) or a different one? I personally have one on my LAN which I use in combination with DDG.
I use https://paulgo.io/ . Works pretty well.
I use searx too, or search within a specific database like Wikipedia or IMDb.
Startpage, Metager and sometimes DDG.
Maybe(to add): https://searx.neocities.org
Policies:
- Redirects users directly to a random selection of any known running server after entering query.
- Requires Javascript.
- Excludes servers with user tracking and analytics or are proxied through Cloudflare.
I found this random Searx/SearXNG search engine redirector at https://searx.space/ under the "About tab.
To use it, I save the search redirector as a bookmark at https://searx.neocities.org/#q=%s&category_general=on
I also turn on javascript on at searx.neocities.org because otherwise I have to type in my search query manually into a specific search instance.
However if your threat model requires it, you can also go to the no-javascript version of searx.neocities.org. You will not be automatically redirected to a random Searx instance when using no-JS version.
I type in a keyword in the url bar and then the service redirects me to a random Searx/SearXNG instance. For load balancing purposes, it seems great since no single server will be stressed significantly. The searches are spread across multiple instances.
I used this in the past but I don’t really like the idea anymore, or at least for librewolf default: the instances are selected randomly and while some things are checked, things I would consider important like the version of searx, are not checked. I also would rather include an instance that has contact infos on the frontpage.
everything but Google really
DDG/Startpage
Removed by mod