I’ve no idea how I overlooked Mojeek. I’m always on the lookout for privacy oriented alternatives. Seems they exist since 2004 and don’t have any controversies surrounding them.
- UK-based
- no-tracking privacy policy
- independent search index
- first search engine to implement a no-tracking policy in 2006
- operates its own web crawler
- infrastructure in a green data center in the UK
- business model based on advertising, API’s and partnerships
I haven’t used Mojeek, so I can’t speak to that, but the UK has some of the worst privacy protections and mass-surveillance anywhere. They’re also part of the Five Eyes, so I wouldn’t count the fact that they’re UK-based as a point in their favor.
Agreed. These weren’t meant to be pros or cons, but facts that I dug up in a quick search. Let everyone interpret for themselves. 😉
Is it something you have to trust they comply with what they say?
Nice that it has its own indexes, but according to this comparison its proprietary SW, running on UK servers without tor interface, and being backed or debated at least by UK politicians. We’re not talking about a not for profit organization either, and they do have individualized answers as well, so they have the mechanisms to individualize results to queries, meaning they keep information about your queries. So in the end, it boils down to the user trusting its service it seems.
Yes, meta search engines do not provide their own indexes, but searxNG is at least open source, you can select the search engines to use, included mojeek, and they serve as a front end preventing the underneath engine to track you (whether it’s against their public policy or not) as if you were to use such engine directly.
this comparison
the table you’ve cited is very out of date when it comes to us and other inclusions
being backed or debated at least by UK politicians
We came up in Hansard yonks ago 2011, mentioned by one MP
they do have individualized answers as well
I don’t quite follow but if you mean results are personalised can I ask where you got that information? One of our main things is that we don’t affect results based upon much more about you than country-level boosts
Perhaps a misinterpretation from mojeek’s wiki:
Mojeek also displays significantly more individual entries in its search results than Google or Bing
Mojeek also displays significantly more individual entries in its search results than Google or Bing
Ah, yeah, that’s from a blogger called Jack Yan who writes a lot on how many results you can actually get out of results 1-10 from n. What he’s saying there, which is correct on checking, is that we will always display 1,000 results when we have them, whereas Google and Bing tend to either stop when you get somewhere in the 200/300s, or just repeat results.
I had used it for some time, but the results were really bad. They have gotten much better, but be aware that you will need to use other search engines. Privacy wise they do collect some basic information such as what you search for and what browser you’re using, but they don’t collect your IP so that it shouldn’t be associatable to you and they never share this info. To my knowledge they have not been involved in any big controversy’s so far.
Got any examples of searches you could do with seeing improvement on? A big part of our process is collecting these (as well as identifying issues in crawling off the back of them)
The searches I have trouble with are mainly support for software. For example, searching for “Kdenlive audio visualizer” results in a Kdenlive tutorial that has nothing to do with audio visualization, Kdenlive release notes and a link to a no longer available Kdenlive docs page. Searching for most stuff with “yt-dlp” will result in the top pages being the alternative to page and some other pages like the GitHub topic that don’t help with the problem at all. While most of the time when trying to get software support you will find a solution, but you will have to search a lot harder. Still, I’m amazed at how much the results improved since my prior use of Mojeek. If it continues to improve like that I’ll definitely give it another go.
Thanks a lot, I’ve raised the Kdenlive query. When it comes to searching for stuff with “yt-dlp” you mean there’s normally another phrase or words in the query? Do you have an example; understand the problem but just trying to raise it with as much of a solid example as possible 🙏
For example “yt-dlp select resolution” or “yt-dlp audio only”. There are some more, it feels like the pages I mentioned are just generally ranked higher.
Also really appreciate the dedication, I think Mojeek is probably the only search engine that listens this much to user feedback. Is there a Lemmy community or something similar where one kann submit queries like these?
Thanks a lot; these have been raised.
I think Mojeek is probably the only search engine that listens this much to user feedback. Is there a Lemmy community or something similar where one kann submit queries like these?
We have a Discourse: https://community.mojeek.com/ but also take these direct from search results pages (there’s a Submit feedback button) or via https://www.mojeek.com/about/contact. It’s all super useful for us so thanks in advance.
What do you use now? Any recommendations?
I made this post a while ago, some parts aren’t correct anymore but it’s still pretty good. I use DuckDuckGo because I rely heavily on its image search filters, but I’m trying to move to a multi engine system. That will likely involve 4get because it’s FOSS and Brave because they have their own index that actually isn’t bad (Brave will be more of a backup solution).
Sounds pretty alright
I use mojeek. It’s good enough for me, image search sucks though
Edit: ups
I sue mojeek
plz do not sue us
The few times I used it it fid a good job of stripping out SEO garbage, but I still prefer duckduckgo for its generally more relevant results.
I just tried it out using the same search string that I’d, just prior, run on the goog:
how many permutations of 9 digit number
Results were not encouraging.
thanks for flagging this, will get it raised
And what is it supposed to tell? I mean, I’d rather ask an llm or some other ai thingy this kind of stuff, and use a search engine for things like looking up docs on calculating the number of said permutations. Besides, there’s no single correct answer to this question, may be 9!, may be 1, may be tons of stuff between them.
There is a correct answer to the question. It’s a math problem.
The task is rather unspecific about what kind of permutations are of interest as well as if there are any restrictions on the number. Like, are 999999999 and 999999999 the same or different?
Yes!
