Search engine alternatives to startpage

[ilu]

2. Search engines: The company behind startpage.com has sold shares to a questionable investor and has been less than transparent about it.

Privacy One Group Ltd has invested in Startpage.com ... Privacy One Group is a separate operating unit of System1, focused entirely on user privacy.

System1 is a adtech company earning their money with data collection. They have already bought other smaller search engines. Privacy One Group Ltd couldn't be found anywhere. Startpage was consequently delisted on privacytools.io .
There has been no proof that startpage in fact changed anything about the way they handle user data. I think delisting startpage at the moment blows things out of proportion considering the alternatives. Duckduckgo is certainly no alternative because it's US-based and we don't know what exactly they do either. Qwant might be considered - did anybody test it? I didn't because it doesn't have the filters (date) I need.

The other listed alternative Searx is certainly the best search option but not for everyone. Searx instances (listed here: https://github.com/asciimoo/searx/wiki/Searx-instances) do meta search and occasionally stop working - not surprisingly they get blocked by the search engines they use or the (privately operated) servers go down entirely or they become unusable for other reasons. And using a search engine named https://suche.honigdachse.de/ might look strange for some - although I can recommend that one (just don't ask me what "Honigdachse" are - probably just an available domain name). I'm using https://search.privacytools.io/ and at the moment it's fine ... but it might not stay fine and thus it can't be our default. And also privacytools.io tracks users with a self hosted matomo script but it gets filtered with uBlock origin anyway (they are open about other ways to disable it too)..

With all the tracking going on elsewhere I think startpage is still ok as default until further notice. People wanting to be more careful should on top of uBlock disable Javascript and cookies for their search engine of choice. Actually, you should do that everywhere. Use uMatrix.

[grizzler]

System1 is a adtech company earning their money with data collection. They have already bought other smaller search engines. Privacy One Group Ltd couldn't be found anywhere. Startpage was consequently delisted on privacytools.io .
There has been no proof that startpage in fact changed anything about the way they handle user data. I think delisting startpage at the moment blows things out of proportion

That was my first impression as well, but after reading privacytools.io's removal issue thread on github and the reddit threads, I must admit, they've been quite patient. Startpage could have prevented this by answering the questions. Not doing so, makes people suspicious.

Duckduckgo is certainly no alternative because it's US-based and we don't know what exactly they do either.

Agreed. I do use it occassionally though and sometimes it gives better results than Startpage. I believe it's Bing based.

Qwant might be considered - did anybody test it? I didn't because it doesn't have the filters (date) I need.

Never used it, or Searx.

With all the tracking going on elsewhere I think startpage is still ok as default until further notice.

Agreed. I'm not sure I really trust them anymore, though...

[grizzler]

(just don't ask me what "Honigdachse" are - probably just an available domain name).

Honingdas: https://de.wikipedia.org/wiki/Honigdachs

[pitfd]

I have removed Startpage as the default search engine
from all firefox installations and moved them over to a searx instance
(it's not Honigdachse ]

to those interested in the topic, have a look at
Martin Brinkmann's of ghacks.net take on the matter

https://www.ghacks.net/2019/11/16/start ... questions/

[ilu]

I've moved the search engine topic to a separate thread to keep the other thread as a to do list. One general remark: I think every search engine and every website owner stores IP addresses for a certain period (default on debian is a week) for technical reasons. A reason why I would never use a static IP for browsing. Important question is: what do they use it for besides fighting DDOS and related stuff?

Duckduckgo: My doubts are referenced here: https://restoreprivacy.com/private-search-engine/. US based, so a no-go.

Searx: The problem about any instance you don't host yourself is that you have to trust the person that's hosting it. As with any other self hosted service. https://searx.me/is the developers instance but it doesn't work for me. Privacytools.io has made a name for themselves in the internet world https://search.privacytools.io and as long as they don't sell out I'm willing to trust them. https://suche.honigdachse.de/ is operated by someone affiliated with Mike Kuketz and if he recommends it thats ok with me. He is also experimenting with his own instance www.kuketz-suche.de. So as long as he doesn't sell out ... and he won't because he's a open source hardliner.

Martin Brinkmann sold out by the way: https://www.ghacks.net/2019/10/07/ghack ... -everyone/ - to a software distributor famous for their shady business modell.

Qwant: French company founded by Eric Leandri and is certainly not financed by by their impressive base capital of €27 714,92 (cheers to publishing that though ). Funded by Axel Springer and European banks according to https://www.cambonpartners.com/en/trans ... stissement. Eric Leandri is also president of https://www.openinternetproject.net that mainly tries to lobby against Google https://searchengineland.com/new-open-i ... ent-191955. Have a look at their member page, you'll find the biggest german publishers (which also means ad companies but mostly from the analog part of the business). F.e. Axel Springer Verlag, who's constantly sueing the company behind Adblock for years now. Qwant fits easily into their anti-Google agenda and their fight pro copyright legislation ("Art. 13" - german users will know what I'm talking about). So Quant is almost certainly financed by advertising companies. And open source is certainly not on their wishlist. All this doesn't mean anything negative for user privacy. I'm fairly certain they will abide by european data protection laws because everything else would compromise their political agenda. There is a political agenda though and it is not mine. But I'm not opposed to considering it.

They have a proprietary but OSM based map service too here: https://www.qwant.com/maps.

Metager https://metager.org/ could be a candidate to replace startpage. Open source like searx but with the instance owned by a non-profit, financed with donations and memberships and advertising. "to receive this advertising, we give the first two blocks of the IP in connection with parts of the so-called user agent to our advertising partners". I think that's well enough de-personalized info. They also operate an anonimizing proxy where no IPs are stored. They have map services under https://maps.metager.de.
I've used it in the 1990es, I should return. They still look the same Anybody else willing to test the quality of their search results?

[pitfd]

Hi ilu,

thank you for the comprehensive description of the
search universe for freedom loving users.

Can I assume that you didn't mention the Yacy project,
because you assume that it is either too complicated
to be set up or is not developed properly anymore?

How would you like to methodically test the practical
suitability of metager?

pitfd

[ilu]

yacy: For the purposes of this forum I'm not considering anything self-hosted. Yacy requires the user to run a publicly accessible service on their machine. This is something I'm not going to recommend to anybody. Ever. I'm running public servers but I'm not going to do that on my desktop. Never ever. This also applies to torrents btw.
There are some public instances but I haven't got anything but "The search service is not available right now. Please try again." Whatever that means. Read https://linuxreviews.org/YaCy. The first paragraph says it all. But read on, it's a good analysis.

Ha, the review is funny, so I'm quoting some: "The search-results are lacking and they appear to be ordered by /dev/urandom." "It loves to just time out." "Leave YaCy running for a while and sooner or later it will decide to peg one CPU core at 100% permanently." "YaCy will effectively behave like a denial of service attack." "The first revision of this page was written in 2007... 12 years later, we have no choice but to conclude that YaCy is a completely useless piece of software." So saaad. (we should really have mrgreen in orange.)
That being said, with yacy you can spider yourself for your own special interest. It's a web crawler, and a very unpolite one. So if you are into that kind of task, that's a very good use case. Just not for a general purpose default browser search page.

Joking aside, searx does a much much better job. If you want to dive into hosting anything yourself, searx should be it. Prepare to constantly fight search provider defense measures though. Let's be honest, searx is leeching off other peoples work. They are leeching our data though, so ... why care?

About testing metager or qwant - just use it. When judging whether the search results match your expectations always remember that you are operating outside of your filter bubble. The search engine doesn't know what you are looking for. As opposed to google, which knows your interests probably better than yourself.

[grizzler]

Startpage replies to questions about ownership change.

[ilu]

Doesn't change anything. Most probably ok for now but doesn't look good for the future.

[Arjen Balfoort]

Lacking transparency is often a way to slowly and silently change policies. Their statement does not reassure me.

So, for the past couple of days I have been using MetaGer and it is not all that bad. For the technical searches I often have to revert to Google but I had to do the same when using Startpage (which is surprising if you think of it).

[Arjen Balfoort]

I have Added MetaGer EN to the list of search engines in firefox-solydxk-adjustments.

This is the current list of search engines:

• duckduckgo
• ecosia
• metager
• qwant
• startpage

We can change the default search engine whenever we need to.

[ilu]

After using it for several weeks now I can confirm that all instances of searx are unsuitable for the average user. I'm switching to metager now to see how it works out.

While searching I stumbled over https://www.wolframalpha.com/ - it mght not focus on privacy but it offers some very specialised search options other engines don't provide, so maybe a nice addition.

[ilu]

I've tested metager as my main search engine for some time now and I'm sorry to say that it's no replacement for startpage. Apart from the mostly worse search result and the lack of possibility to filter for "last year" (which I use regularly) I also dislike their habit to route search result links (randomly?) with an ID. I can't reproduce that today but I saw it on and off till yesterday.

Edit: It's not random. It's https://www.smartredirect.de/redir/clickGate.php for "partnershop" and it's done whenever there is a logo (uMatrix catches it though). I can understand the reason, they need money, but I think non-tracking advertising that startpage promises to do is better. If they keep this promise.

I'll switch to quant for bread-and-butter searches now and see how that works out. But whenever a search gets difficult I'm still using startpage. It has the best filters and thus the best results

[pitfd]

Hi ilu,

interesting - thank you for sharing your experience
with MetaGer.

I can hardly believe your discoveries, as the search engine
is run by SUMA-EV - Verein für freien Wissenszugang (e.V.)
in cooperation with Leibnitz university of Hannover. Their
data protection policy is laid out here:
https://metager.de/en/datenschutz/

Among other things, they write:

quote on:
User Agent Name

When a website is opened, your browser automatically sends an identifier, usually with data about the browser and operating system used. This browser identifier (the so-called user agent) can be used by web pages, for example, to recognize mobile devices and present them with a customized output.
Example
Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:59.0) Gecko/20100101 Firefox/59.0

What does MetaGer/SUMA-EV do with it?

In addition to donations and membership fees, we must finance our operation through non-personalized advertising on the results page. To receive this advertising, we give parts of the user agent in connection with the first two blocks of the so-called IP-Adresse to our advertising partners.
quote off

Can your findings be reproduced and if so, how? I run noscript and
I did not see any other server beeing contacted.

BTW: I have changed to use one of the few searx instances
beeing around, my fallfack solution is startpage.

pitfd

[ilu]

Yeah, I know. I was wondering myself. Search on metager.de for something like "buy refrigerator" and have a look at results sporting an icon (which are a "Partnershop"). You'll see the link goes to api.smartredirect.de with an ID - whatever is encoded there I don't know but it doesn't look like a personal id. Upon closer investigation this happens as soon as I use metager.de and the corresponding search plugin. It does not happen on metager.org. This explains the on-off nature.

The search plugins (https://metager.de/plugins/opensearch.xml and https://metager.org/plugins/opensearch.xml) are not different but the server seems to be.

There's another kind of link like https://metager.de/r/metager/<very long number> - the first part of the number is always the same for me. The last part somehow encodes a search result link. I can see this in my history and reopen the link from there, weeks later.

(note: I'm running with JS disabled. uMatrix catches the redirects for me. Maybe that's why I can see them in my history.)

Whatever this is, they need to explain it in their privacy declaration. I thought about contacting them (they will come back to you after some time), but I could not find the time.

I'll be focusing on metager.org now.

In case anybody is interested in investigating search plugins: https://www.jeffersonscher.com/ffu/searchjson.html