So with the recent Bing situation I wanted to take a second look on private search engines and sharing my conclusions of each search engine. Here is my list of private search engines:
I really like Duckduckgo, it has all important tools, decent result quality and a great image search function. Instant answers is very useful. My main problems are the reliance on Bing as the index and the choice of Apple Maps as mapping solution. Apart from the situation with the browser and Microsoft tracking Duckduckgo has a pretty clear record and the privacy.
Startpage is another great option. Apart from mapping everything is there and, while not as good as Duckduckgo's, the image search engine good. The results are based on google and on par to better than those of DDG. The main advantages over DDG are European base (Netherlands) and the anonymous view, which basically functions as a quick access VPN, but sadly breaks ad/tracker blockers. Privacy for regular search is equal to DDG, but you have to disable JS to get rid of some telemetry. It is owned by an advertising company
Swisscows is okay. It is also Bing based, but slightly worse than DDG results. It lacks image search filters and mapping, but offers a music search which allows you to listen to ad free music. It also has an anonymous view, but it's not interactive. Privacy is similar to DDG, but has more telemetry and (temporally) stores your IP. It is from Switzerland, it also has a very strict anti gore/porn policy that sometimes makes normal search terms inaccessible.
Qwant used to be very solid French search engine, has dropped in quality. Similar search quality to DDG, image search like Startpage. They use Bing in combination with their own index. Then problems: They share your IP with Microsoft and they replaced their main advantage, openstreetmap based independent mapping service, with AI summary's that require an account. Worse privacy than all the above.
Very similar to DDG. The main differences are that Ecosia is based in Germany, it plants trees to fight climate change, but also forwards your IP to MS.
Braves main advantages are being independent, both with the search and the AI, and the goggles that allow you to customize your results. Search results are slightly better than DDG, image search is bad, no mapping is available. Brave has had invaded privacy in the past, but currently the privacy is good as long as you disable statistics. The company itself is a bit concerning and the CEO is homophobic.
SearXNG is self hosted and open source, it uses various search engines as index and has a ton of extra feature like music search, fediverse search and a bunch more. While it has the most features and best privacy of all options, public instances are sometimes slow and the results aren't really good.
Kagi is in principle a decent quality search engine, but it is paid and has some problems that are only getting worse. For those interested read this blogpost.
4get is a open source, self hostable search engine. It acts as a web scraper for various search engines, also supports Soundcloud. It has great privacy and good results, but it lacks mapping and the official instance requires a CAPTCHA per 100 searches
Gigablast Open source, self hostable, independent search engine
Mwmbl Open source, independent, self hostable search engine. Only web results
Marginalia Open source, independent, self hostable search engine. Only web results, offers filters
That would be my list. I'll still be sticking with Duckduckgo but I'd reconsider if Startpage improves it image search. Brave will probably never be my default, but it has proven it's role as a more private backup. Comment if I missed any search engine
thank you for linking this! It made me aware of stract.com, which seems to be a wonderful project. I hope it grows and gets better so I can use it as my main search engine next to Searxng one day
I thought that Kagi would have way more users. That blog was an interesting read. If that is their financial management, they're doomed to fail. The founder also seems somehow worse than Brave's. But it does give me a chance to mention something I've been thinking about for the past 6 months.
There's right now a massive trend towards co-opting in tech. Where startups and corporations use current trends in the tech savvy consumer to push products and services that ultimately actually go against the trend. Privacy, security, federation, climate change, open source. But just like most con men, it's all performative, not substantial. They are trying to get fast to the wallet, then run for the hills with it. It reminds me of common greenwashing from oil companies, I call it privacywashing. In the end they still get to keep your data, and push anti-consumer tech like blockchain scams and fraudulent AI.
I saw an extension the other day that let you choose which mapping service you want to use when clicking a link to a map. You could click on a Google Map link but have it open in OpenStreetMap. If I find it again I’ll update here
What is this?
This is a metasearch engine that gets results from other engines, and strips away all of the tracking parameters and Microsoft/globohomo bullshit they add. Most of the other alternatives to Google jack themselves off about being ""privacy respecting"" or whatever the fuck but it always turns out to be a total lie, and I just got fed up with their shit honestly. Alternatives like Searx or YaCy all fucking sucks so I made my own thing.
Looks promising. Do you know if 4get supports bangs (like in duckduckgo, for example !nixpkgs lemmy searches for Lemmy on Nixpkgs' website)
The developer wants to add redirects to frontends in the future + also a token which let's you use the main instance without needing to solve a captcha
Wow that's quite the thorough list. You've included a bunch I've never even heard of. I use a combination of public SearXNG instances, DDG, and Leta (Leta acts as a proxy for Google and Brave search results and is only available for Mullvad customers.)
I just switched to Kagi because I liked the idea of a paid search engine who's aim was to remove the internet's clutter, not use any profile besides the one I create to show me results, and where I could weight certain sites that produce good content.
Reading the blog post the issues allegedly are:
Privacy is not guaranteed, like with a 3rd party audit
AI usage is growing not shrinking
The business seems to be poorly run and could have a short lifespan
Yeah that blog post didn’t contain anything damning, just generic level “business could be run a little better”, and “they like AI more” than the blogger dies. Ok, so, and? Whole lotta smoke there without much fire. I still am happy with Kagi from what I read.
The fact that much of the project is discussed over Discord should be damning enough...
Don't you think is damning that someone's concern of privacy is handwaved because the data was "volunteered, not collected" ?
And, as a user, many of the AI features are limited at best and factually incorrect at worst. I would only salvage FastGPT and Quick Answer; they summarize the first 4 or 5 relevant links and contrasting views (even if it bases the whole search on a single Reddit comment from 2017).
Funnily enough the Universal Summarizer and Discuss Docs are the least reliable.
All in all, I am conflicted because it seemed like a pristine service and it's getting clouded with time.
I used to use them a lot before the system1/Privacy One acquisition, but then decided to switch back to google (never liked DDG's results, but might have to try them again).
This is a very comprehensive summary, thanks for the effort you clearly put in.
If I can make one correction, it would be to clarify that Ecosia is not really comparable to DuckDuckGo in terms of privacy. Not only does it log your IP address, but it also logs your search queries and forwards both of these to Microsoft and/or Google (depending on how you choose to search). Ecosia anonymises your IP address after a week, but for Microsoft that process takes 6 months and for Google it takes 9 months. In contrast, DuckDuckGo does not log your IP address and only collects anonymised search results, completely separated from any personal identifier. It does not forward this data to any third party. DuckDuckGo has also made privacy-guaranteeing agreements with Microsoft around ads (which are provided by Microsoft). Ecosia has not made similar agreements with Microsoft and Google from what I can tell.
Thanks for the linked blogpost about Kagi! That saved me a lot of money.
It really changed my mind about it.. I want to support a privacy focused search engine, not an AI-loving mess without focus on their core product.
I will have a look at MetaGer next. Although I am sad that there is still no good original search engine with it's own results.. Still all depending on Google and Microsoft.
Ill check startpage out! Also an interesting blog post regarding kagi. Ive never used it in the past, but now i definitely wouldnt ever even consider it.
The software itself doesn't require any Captcha, it's just the specific instance you chose that has this requirement. There are many other instances (like e.g. 4get.ch) that don't have a Captcha at all. You can find the entire instance list at https://4get.cynic.moe/instances. It also shows you whether an instance has the captcha enabled or not.
Maybe not the best wording (English isn’t my native language). Brave collects large amounts of data by default, but the problem is/was mostly in the browser Read more
I stopped using Brave for anything serious after diving into the data collection of Leo, the built-in, practically unstoppable webbrowser AI. It's appaling to me. I use Cromite at the moment.
I used to be a Kagi subscriber because I believed in their image for Orion. Their strong views on privacy, imo, directly conflict with their action to keep the product closed source “because it’d slow them down”, so I ended up unsubscribing. Good to see I unsubbed just in time.
Like, I vaguely understand the terms everyone uses to explain it but I don't really understand what it does or how it does it. I've used a public instance of it that the maintainers of my Linux distro provide and is set as default search on a fresh install. The results weren't terrible but did take some time to load, which is the main reason I tend to use other engines.
If I self host it do I get better performance? What about results? Are they different on different instances?
SearXNG searches google, bing etc. for you and shows them to you. Speed depends on server speed and user number. Even if you don’t self host you can change the search providers in the settings, which can have an effect on result quality, but the less you choose the faster
It didn't return any search results during the Bing outage, as experienced by me personally and reported by others. At least one blog post claims it relies on Bing.
I have been really wanting to try out Ekoru but it just doesn't work for me. When I search something with it it just gives me the search bar and menu at the top but it doesn't actually show searchr results. I have tried it with VPN off, different DNS, non hardened stock Firefox, and still couldn't get it to work. Can anyone else confirm if they can use Ekoru or not?
I have tried all the engines for this post and Ekoru has been working fine. I’m using the Firefox rpm on Fedora, slightly hardened, no VPN, Quad9. Make sure the Ekoru extension is up to date and has all the necessary permissions or alternatively try lilo
Yea I've tried it on Firefox flatpak for Linux and also Firefox on Windows, downloaded the extension from the official Firefox Addons, and nothing worked. Very strange. I probably won't use it anyways as it seems very lacking and redirects you to Google and Microsoft for image and video searches which Ecosia doesn't (it's what I have been using for almost a year now). I just thought cleaning oceans was a better cause but oh well.
Does having Google as index puts the search engine inside a Google Bubble? Like if every anonymous queries made by that engine sent to Google makes a big profile used by Google to aim results or do they use some type of API avoiding this?
I stopped reading after Brave, as you chose to derail your product reviews to meddle in someone's personal beliefs. Those two things have no correlation.
If someone using Brave gives him money and that money goes in to a homophobic lobby it would be better for consumers to know that so they can actually consent to that. Consumers deserve to make informed decisions about who to or who not to support.
Sure, consumers should make informed decisions. But this was a technical\feature review until OP brought in their personal feelings on someone's personal beliefs. Once they did that, they no longer were doing technical review on search engines.
someone being a dickhead givs me the right to say im no t sure about their product, bc theyre a dickhead mrow.
this is good journalism, as it shows not only where the product is, but where it may be going (and a link so u can see for urselfs).
OP dids a very good job
if u want me dead i havethe right to tell u to go fuck urself :3