c/selfhosted · by mesamunefire@piefed.social · 252d

Be Your Own Privacy-Respecting Google, Bing & Brave

https://xn--gckvb8fzb.com/be-your-own-privacy-respecting-google-bing-brave/

… by running your own instance of the free and open-source federated metasearch engine SearXNG on OpenBSD!

222 points · 47 comments · view on lemmy.world

47 Comments

yodeljunkmanenvy@piefed.social · 30 pts · 252d (7 replies)

There is also a list of publicly operated SearXNG instances at https://searx.space/. We host the one at https://search.freestater.org/, and there are plenty of other good ones.

Appoxo@lemmy.dbzer0.com · 3 pts · 251d (5 replies)

Who says they are securely operated and don't store any data??

jackr@lemmy.dbzer0.com · 7 pts · 251d (1 reply)

who says that about any search engine? can you trust them? searXNG is usually run by random people who are less likely to use your data than a larger search company

Appoxo@lemmy.dbzer0.com · -5 pts · 251d

At least with Google it wont be an individual that can open my data (most likely).
Probably only the AI or whatever can actually read any of it directly.

yodeljunkmanenvy@piefed.social · 3 pts · 248d (1 reply)

The devs at SearXNG have a bot that regularly scans the public instances for changes to the source code and delists them as a public instance if it's altered.

If you go to https://searx.space/, they show the results of the scans for each instance.

The software is free and open source. You are encouraged to inspect the code yourself to make sure no data is collected!

Here is the source code:

https://github.com/searxng/searxng

Appoxo@lemmy.dbzer0.com · 1 pts · 248d

Thank you very much for proving what's correct :)

quick_snail@feddit.nl · -5 pts · 251d

If you're using a shared IP, it doesn't matter.

You are using a VPN or Tor, right?!?

yodeljunkmanenvy@piefed.social · 1 pts · 248d

[ removed ]

thejml@sh.itjust.works · 24 pts · 252d

I run it in a container on Kubernetes. Definitely recommend.

jol@discuss.tchncs.de · 18 pts · 252d (3 replies)

I still don't understand how Searx is able to operate for free. Don't the API calls cost money?

irmadlad@lemmy.world · 19 pts · 252d

From what I've read, I believe it's a combination of donations. sponsors, volunteer hosting from like minded organizations.

Neptr@lemmy.blahaj.zone · 9 pts · 252d (1 reply)

If I had to guess, they probably don't use the APIs, inside using scrapping of some sort.

sunstoned@lemmus.org · 1 pts · 251d

My understanding is it scrapes what it can't meaningfully get out of an API. Public instances run into rate limiting, but private instances don't really have that problem.

ap1werks@sh.itjust.works · 16 pts · 252d (2 replies)

look at that domain name! respect

Cyberflunk@lemmy.world · 4 pts · 251d

[ removed ]

quick_snail@feddit.nl · 2 pts · 251d

Usually its a sign of a scam. Looks like a ransomware c2 domain

carrylex@lemmy.world · 12 pts · 251d (4 replies)

For anyone wondering xn--gckvb8fzb.com is マリウス.com

Also: Maybe they should get a "normal" domain when you post english articles...

mesamunefire@piefed.social · 1 pts · 251d (3 replies)

Does the url not resolve on certain browsers or something? I usually just copy/paste or use a firefox plugin to generate posts for lemmy/piefed/fediverse.

carrylex@lemmy.world · 6 pts · 251d (2 replies)

It resolves well but punnycode is disabled in some browsers or security enhanced profiles, so that you can easier detect punnycode domains that try to fake other domains.

mesamunefire@piefed.social · 1 pts · 251d (1 reply)

Im not familiar with punnycode. Interesting.

sludgewife@lemmy.blahaj.zone · 1 pts · 251d

it's punycode

grue@lemmy.world · 7 pts · 252d

Is metasearch really the best we can do? What about YaCy, or something else more like that?

pineapple@lemmy.ml · 7 pts · 251d (5 replies)

The search engines that searxng interact with still track you. For this reason I will always use a public instance to mix up the tracking with everyone else using it.

TractorDuffy@lemmy.world · 0 pts · 251d (4 replies)

Explain?

pineapple@lemmy.ml · 3 pts · 251d (3 replies)

Using a public instance is more private than using a private selfhosted instance.

TractorDuffy@lemmy.world · -2 pts · 250d (2 replies)

..a public Lemmy instance? How is that related to the the search engine?

044en17rl@lemmy.world · 2 pts · 247d

a public SearXNG instance

pineapple@lemmy.ml · 1 pts · 246d

044en17rl is right I meant a public seaxng instance.

biotin7@sopuli.xyz · 7 pts · 251d (5 replies)

Or how about YaCy. It's self-hostable & you can have your own web index and start your own web-crawler.

It's peer-to-peer too

motruck@lemmy.zip · 3 pts · 250d (1 reply)

This is exactly what I came here to find. Thank you for posting it. If I can be so bold selfhosters should really be leaning this way searxng is great but it still uses big tech.

The other thing we need is a way to identify good crawling agents or *smol agents over corporate bots that just steal content.

If selfhosters can unite and build a good index perhaps searching can go back to the way it was vs a vector to sell you more and collect your data.

biotin7@sopuli.xyz · 1 pts · 250d

You know what else I like ? Fossil

It's a FOSS alternative to Git itself, their official site itself is a self-hosted instance of Fossil.

044en17rl@lemmy.world · 2 pts · 247d (1 reply)

How good are the results compared to Google/Duckduckgo?

biotin7@sopuli.xyz · 1 pts · 246d

You make your own index with this one

mesamunefire@piefed.social · 2 pts · 251d

I personally love yacy.

N0x0n@lemmy.ml · 5 pts · 251d (9 replies)

I used to self-host searxng for a while, but somehow the search results were always off and mixed with to much non-relevant results :/.

It's not about searxng itself... Rather how the most relevant info gets drown into AI slope and non-sense bullshit. The best blogposts/info are transmitted from people to people...

I'm kinda sad to admit that stupid AI "solved" this issue and had better results :/

A_Random_Idiot@lemmy.world · 3 pts · 251d (3 replies)

I used to self-host searxng for a while, but somehow the search results were always off and mixed with to much non-relevant results :/.

I mean, getting non-relevant results happens with every search engine anymore.

The days of your search results being relevant, and what you want on the first page, are long dead thanks to SEO and other factors.

N0x0n@lemmy.ml · 2 pts · 251d (2 replies)

Yeah you're right ! However, ages ago, I still remember how you could go to page 20+ and still find some really interesting things !

Here, past page 2 it's just some random shit...

A_Random_Idiot@lemmy.world · 2 pts · 251d (1 reply)

Most of its just AI generated websites. Search for any topic and you're likely to get 50 AI generated website that give a similar bulletpoint presentation of what search item X is according to the AI that generated the site.

N0x0n@lemmy.ml · 1 pts · 251d

50 AI generated website that give a similar bulletpoint presentation of what search item X is according to the AI that generated the site.

Yeah that's what I meant by "random" shit ! 😁

sunstoned@lemmus.org · 2 pts · 251d (3 replies)

You can self host that too ;)

OpenWebUI + Ollama + SearxNG. OpenWebUI can do llm web search using the engine of your choice (even self hosted SearxNG!). From there it's easy to set the default prompt to always give you the top (10, 20, whatever) raw results so you're not confined to ai results. It's not quite duck.ai slick but I think I can get there with some more tinkering.

dubyakay@lemmy.ca · 3 pts · 251d (1 reply)

Is there a guide on how to do this on Linux + 16GB Radeon?

sunstoned@lemmus.org · 7 pts · 251d

I mean, I could write one! I kind of just pieced it together from guides on the three individuals

Edit: back of the napkin guide below is basically in the OpenWebUI docs already! I use NixOS (btw) but docker/podman should work well.

OpenWebUI + Ollama setup -- tl;dr docker run -d -p 3000:8080 --add-host=host.docker.internal:host-gateway -v open-webui:/app/backend/data --name open-webui --restart always ghcr.io/open-webui/open-webui:main

OpenWebUI SearXNG guide -- a little more involved, but not difficult.

N0x0n@lemmy.ml · 1 pts · 251d

Ohoho? That's interesting. I don't have the horse power to selfhost an AI, but that's good to know !

Thanks for the pointer !!!

Fleppensteijn@feddit.nl · 1 pts · 251d

[ removed ]

MentalEdge@sopuli.xyz · 5 pts · 252d

It's not federated tho?

What do they mean when they call it that?

ArchEngel@lemmy.ca · 4 pts · 252d (1 reply)

Thanks for posting, both a great reminder to try setting this up on my unraid, and also to add the RSS feed of that site to Feeder.

pipes@sh.itjust.works · 3 pts · 252d