Every search engine will use your location. Especially when you make location based searches.
Do you think “restaurants in my area” should mean Earth? Or would you include the ISS also?
Ideally it should also include the restaurant at the end of the universe
How is that place still in business
A place like that relies on repeat customers
Any website you connect have access to your public IP for each request, hence location recommendations
Content deleted by creator due to lemmy.ml tolerating brigades from hexbear
lol
Content deleted by creator due to lemmy.ml tolerating brigades from hexbear
You use the word “harvest”. Which has a meaning along the lines of: Collect and save for later use. That’s not necessarily the case in your example though. Nothing needs to be collected or saved. It only needs the one IP, the one time that search is done, then it can be (and on most of your “tested” engines is) forgotten.
If you search for “restaurants in Edmonton”, you’ll likely receive exactly what you think.
If you search for “anarcho-communism” or something equally non-related to location. Than the IP won’t matter, beyond giving you results in your local language.A truly blind search, would generally suck. You’d need nearly half a dozen qualifiers to find some relevancy.
Content deleted by creator due to lemmy.ml tolerating brigades from hexbear
That’s just a test though, to see if the search engine is using some form of tracking. If they searched for something like “news” or “hot milfs” they don’t want “hot milfs in your area”.
That isn’t tracking. That’s just using the IP the search came from to specify “my area”.
No one said it was tracking. Is using location as a search term without it being asked for. What OP is looking for is a search engine that would return different results for “restaurants” and “restaurants near [location]”
Content deleted by creator due to lemmy.ml tolerating brigades from hexbear
No search engine can be neutral, for the same reason that no news can be neutral.
People must decide what it will show you and if an algorithm decides, people must program the algorithm. All people are biased in some way.
True, but there is no requirement for a search engine to use such an algorithm to net results. I know this because I’ve been using search engines since the Windows 95 era. It’s possible to have what I seek. I’m just starting to question if anyone makes it.
I’m going to try the .onion version of DuckDuckGo later and see how that goes.
Update: This is … extremely odd. Simplified explanation: Tor connects you via a specific circuit of nodes, which you can see by clicking an icon next to the address bar. The last node it connects to is the exit node. But here’s the thing: Even the Tor edition of DDG insists on providing location-biased results that match the location of the exit node. I tested this by running a search on DDG and then connecting via a different circuit several times. Each time, the results were tailored to the location of the exit node. This is very disappointing. I’m going to do some searching and try some different search engines, but I think it’s safe to say I will not find what I’m looking for in any fork of DuckDuckGo, not even the .onion version.
Update 2: Ooops, I was using the clearnet version. I switched to the .onion version, which it turns out has to be done manually. I’m still getting very different results based on the exit node, although it’s much less obvious now. I’m thinking there might be some other .onion search engine that does what I need.
Not sure it will solve your problem, but if you’re not happy with the public SearXNG instances, you can run your own instance of SearXNG on your local machine, and even set up custom filters and redirects to get rid of SEO junk.
Content deleted by creator due to lemmy.ml tolerating brigades from hexbear
If you want to give it a shot I would recommend you use Podman Desktop. What I did was I just followed the instructions from the SearXNG page until I had to run it. Afterwards I would just add it to Podman Desktop. Then enter the settings to set a localhost. You can use 8080 in the localhost setting. After that I would just press run and it would start. On firefox or whatever browser you want, now enter
localhost:8080in the search bar and you have your very own searxngContent deleted by creator due to lemmy.ml tolerating brigades from hexbear
Use something like tailscale if you want to easily and safely use it over the web (IP will then be 100.something something)
deleted by creator
I see you already have an answer using podman.
But don’t be afraid of the command line. If you can copy/paste a few commands, it’s pretty easy to set up.
I honestly find installing docker harder than to start a locally hosted searxng instance.
Also, something like self-hosting your own email is way harder and requires a lot more maintenance. I’d leave that project to further down the line.
Content deleted by creator due to lemmy.ml tolerating brigades from hexbear
Duckduckgo has a toggle to disable location-based results though? You can also choose a different location if you want to.
Content deleted by creator due to lemmy.ml tolerating brigades from hexbear
This one. I haven’t tested it systematically but it works well enough for me. If I search just for ‘restaurants’, the map will use my location, but the regular search results are tripadvisor and opentable results for Chicago, Bronx, and Milwaukee, none of which are remotely close to me.
Edit: Just realised the marking wasn’t saved, but the toggle is right above the map.
Content deleted by creator due to lemmy.ml tolerating brigades from hexbear
I think the map part is separate, the regular results didn’t take my IP into account from what I can tell. But I can’t guarantee that of course. Otherwise you could also check out mojeek. From their about page:
We act on our own agenda and not that of others, this is why we focus our time on one thing and one thing only - “No tracking. Just Search.” When you conduct a search on Mojeek, your results are based entirely on the keywords you typed in. Mojeek does not possess any previous identifying information on you, such as IP addresses, search history or click behaviour.
Content deleted by creator due to lemmy.ml tolerating brigades from hexbear
I think Qwant is as close as you’re going to get unless you can set up searchxng to do what you are asking.
Now, you might be able to get better diversity in results if you use a vpn to move to more diverse or contrasting cities.
I often find news sources external to the US to be very interesting insights to what we see rammed down our throats.
Content deleted by creator due to lemmy.ml tolerating brigades from hexbear
So many of the engines you mentioned by default geo-locate you for search relevancy, but you can turn that off. I believe Qwant, DDG, and Kagi all have configuration settings for that. Generally what you want is what is sometimes termed the “international” edition.
However, that being said, you’re never truly pulling search results from outside the anglosphere because you’re entering search terms in English.
I know you mean well, so have an upvote. It’s true there is a setting in DuckDuckGo to select a region, or to leave it set on “all regions.” Unfortunately, it is functionally useless since I’ve confirmed that it does provide data tailored to my IP. In fact, not just my IP, but the name of my exact county and home town. The HTML fork does not have such a setting, but it was only more subtle about providing location-tailored results. At the end of the day, I can’t rely on it.
Qwant doesn’t have such a setting, but at least it gets as general as “United States” (which would be fair enough for me – like you said, and it’s also occurred to me about searching in English – but searching English terms is also deliberate input from the user and would fall under the category of “keywords and syntax.”) However, Qwant does not simply provide US-based results, but actually targeted toward my exact location. I’m not aware of any setting in which I can change this. If you know of one, please let me know where to find it, because it’s not in the regular settings page.
Kagi is the same as Qwant, but with the addition of an “international” setting. Again, it is non-functional since it persists in providing very specific results to my location. In fact, Kagi is one of the worst I’ve tried so far. It even presents the results in a style reminiscent of Google, including starred reviews and all.
I find it ironic that I avoid toxic algorithms like the plague by not using Facebook or Twitter, yet I am seeing data manipulated by the very search engines I use every day. There’s got to be some obscure old-school search engine out there that works like they did in the 90s, right?
I just learned about kagi. It’s subscription-based but seems really nice
Content deleted by creator due to lemmy.ml tolerating brigades from hexbear
On one hand they are incentivized to not f-over their users. On the other hand, because you need a paid account it wouldn’t be as private as SearXNG-over-Tor or whatever.
You Will NOT Find A Search Engine That Does Not Geo-Locate You. They DO NOT Exist.
Why am I so bold in my statement? Because they don’t exist. Please oh please try to prove me wrong, it will be very entertaining, and I promise I will find that every engine you recommend will be caught red-handed doing this by the time I complete 100 searches specifically crafted to bait this behavior out.
How do I know? Been accessing the public internet since 2004. They all have been doing so since then; and those who failed to do so have ceased to exist.
How do I evade it? Unfortunately, you don’t I recommend using either Tor; or a VPN. Then you’ll know what region and possibly city your accesses will appear from; and the blatantly localized results will be irrelevant to you.
But XYZ has an option!~ No, they do not. You will still receive data relevant to your language and country as determined by your IP Address’ Geo-Location. You can’t turn that off; and engines won’t give you the ability to ignore fine-grained IP location either if you ask for something local; which still localizes you to the city level.
Geo-Location is a core feature of all search engines. So good luck trying to avoid it.
Content deleted by creator due to lemmy.ml tolerating brigades from hexbear
Onion Search engines are depending on the Search engine what results they provide DDG’s Onion Site provide Clearnet search results like the normal one. You could also try LibreX and Brave Search.
Nice. After work today, I plan to download Tor, and I’ll try these out! Iirc DDG is the default search engine in Tor, so that should be the first I try!
It will be a bit of a pain having to open Tor to search, but since it would just be for searching I don’t think it will be too complicated. Do you think there’s a way to set Tor to open clearnet links in normal Firefox? That would be ideal if I could search within Tor but open the links in my regular browser.
Not sure, there is maybe some extension for this, I would suggest to do some search and try to find smth, but you could also open clearnet links with Tor, it’s even more private that way if you do it right.
Yeah true, but some sites do break in Tor. I’m not that concerned about some specific sites. I just want control over the choices that I make instead of having a search engine try to do the thinking for me.
deleted by creator
Depending on what you’re doing, Local LLM can help a bit. Like if i want a recipe for an apple pie i could use LLaMA-2 to find out even without an internet connection.
Not saying its a replacement for a search engine, i just think its worth mentioning.
(edit for grammar)
Content deleted by creator due to lemmy.ml tolerating brigades from hexbear
