Have you looked at this ?

https://seirdy.one/posts/2021/03/10/search-engines-with-own-indexes/

This doesn't just cover what the title says, it also shows who uses what indexes. It's the most comprehensive list I have seen.

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.

https://www.privacyguides.org/en/search-engines/

So are there search engines that use open street maps?

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:

1. Privacy is not guaranteed, like with a 3rd party audit
2. AI usage is growing not shrinking
3. The business seems to be poorly run and could have a short lifespan

Is this correct?

There's also 4get (metasearch engine; https://4get.ca is the developer's instance).

Thanks for the list!

One point I will add is that Startpage is actually owned by an American spamming company:

https://restoreprivacy.com/startpage-system1-privacy-one-group/

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).

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.

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.

Startpage recently started collecting browser/OS telemetry using JavaScript so if you want to use it you should disable JavaScript on their site

I would use kagi, but only if they supported monero payments.

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.

4get

requires a CAPTCHA per 100 searches

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.

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.

Can someone link me the Brave privacy violation? I’m a current Brave browser and search user and prefer to stay informed.

I mainly use Andisearch, maybe one of the most privacy protecting search engine with AI. It's PP is pretty impressive

Thanks a lot for the write up!

Can someone explain like I'm 5 searXNG?

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?