Search Engines

I use several, duckduckgo, noai.duckduckgo, Bing which is most correct lately, Google, all depends, there are too many out there.

I strongly do not recommend using any DuckDuckGo services. Not only do they openly admit to not removing Microsoft's trackers, but they're actually engaging in even more malicious non-transparent tracking: for certain specific search queries or browsing modes without JavaScript and cookies, they include referrer HTTP headers similar to Google: https://duckduckgo.com/?uddg=something_they_want_know_which_link_you_click Additionally, for Tor Browser, there are undisclosed restrictions: when accessing html.duckduckgo.com, they test your browser's JavaScript and cookie capabilities (although they don't execute anything now, but who's to say they won't cooperate with hackers to inject a web 0-day vulnerability in the future?). If your browser lacks the execution capability, you'll be redirected to the Tor onion version of DuckDuckGo, and the Tor version requires enabling JavaScript and cookies, effectively soft-forcing all Tor users to have JavaScript execution capabilities.

Searx is also not recommended. In their GitHub issues, the development team stated that if JavaScript helps with development, they wouldn't remove JavaScript dependencies. SearXNG is better, but as you mentioned, the problem is there simply aren't any stable instances.

For Tor users, I personally recommend Startpage's onion version, which has no JavaScript requirements and no client-side verification. However, their clearnet version has added JS-worker verification, and the captcha URL also uploads your country, ISP, and other detailed information for their tracking: https://startpage.com/sp/captcha-block?bc=US&bi=Surfboard+Holding&be=[id_and_language_info+search_query+date+search_method] Therefore, I have no clearnet search engine recommendations and welcome others to share theirs.

@GlitchFrame that would also depend on which browser, plus which OS you use.
There are headless browser too

@NickZ2 IDK which part you're referring to, but it doesn't matter. A good privacy search engine shouldn't force users to enable JavaScript. Most of the web fingerprinting techniques I mentioned cannot extract much information without detailed JavaScript feature exploitation. The web server itself can only utilize User-Agent, HTTP version, and ACCEPT-Header. and almost 90% of identification relies on User-Agent. Most people who use ddg or startpage know at least how to modify their UA.

As for headless/terminal/TUI browsers, I typically use elinks with CURL backend. By modifying headers and request methods, it can send headers identical to Tor Browser in safest mode: HTTP/2.0 "gzip, deflate", "text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8". In the web server logs, there's no way to distinguish any difference.

Of course, theoretically one could detect the difference through TLS JA3 fingerprinting, since the TLS fingerprint differs from Tor Browser's. However, to my knowledge, hardly any service uses such deep-level user identification."

@Stylus1063 it's really not.

Startpage is the best for commercial users...but it limits thee to 100 searches.

Searx ENG is da shit but you gotta set dat shyte up.

I’m surprised no one has mentioned Kagi.

It’s not free, and that keeps a lot of people from trying it. But after about a year of using it, it’s worth every penny to me.

@Geek Kagi sucks simply because they were going to add Monero support and told people they were going and then decided against it. Beyond frustrating.

@GlitchFrame 1st of all not everyone wishes to hide anyway.
The privacy debate is exaggerating many risks.
I use a lot of different browser to check.

@xenu Agreed, but it's still the best option as of writing in my opinion.