In terms of privacy, this is how the Searxes (meta of meta searches) compares to DDG, Startpage, and Mojeek:
| privacy factor | DDG | Startpage | Mojeek | Searxes |
|---|---|---|---|---|
| caught violating privacy policy | yes | no | no | no |
| bad track record (history of privacy abuse) | yes (CEO founded Names DB) | owned by targetted ad agency | no | |
| feeds other privacy abusers | yes (Verizon-Yahoo, Microsoft, Amazon, CloudFlare) | yes (Google, CloudFlare) | no | no |
| privacy-hostile sites in search results | yes | yes | yes (but appears less frequent than ddg) | no (CloudFlare sites filtered out) |
| server code is open source | no | no | no | yes |
| has an onion site | yes (but Tor-hostile results still given) | no | no | yes |
| gives users a proxy or cache | no | yes (using Anonymous View feature) | no | yes (via the favicons) |
Superficially Metager is privacy respecting and there’s even an .onion host for it. So I’ll have to add it to the table in the future.
For the moment, I’ll say that Metager shares the following with advertisers:
- first 2 blocks of your IP address
- user-agent string
- your search query They say it’s for non-personalised advertizing.
deleted by creator
Mojeek does their own crawling. That’s quite impressive, because unlike DDG it means they don’t have to choose from a pool of privacy abusers to buy search results. I did a test search on “petition sites” and was impressed that the first page did not contain the typical privacy abusing cloudflare results (change.org, moveon.org, etc). Mojeek also does not buy hosting from Amazon/MS/Google. I should perhaps add them to the table.
I didn’t know MetaGer was free software. I occasionally use metager.com but my first port of call is Searxes b/c searxes filters out cloudflare sites.
deleted by creator
deleted by creator
StartPage recently was acquired by an ad agency
Good find… I’ll add it.
Also not really sure why you would include a “tor-hostile” and “privacy-abusers” websites included in the web search included section… censorship, any kind of censorship is bar regardless…
Can you clarify what you mean? The note about Tor-hostile results is on the row labeled “has onion site”. Generally, only Tor users visit onion sites. So if you give a Tor user links that they cannot interact with (e.g. b/c they’ll get a 403 access denied error or broken CAPTCHA), the content is already censored by the webmaster. Polluting the search results with such broken/unusable links actually suppresses links that would be more useful by pushing them down.
Note as well the Searxes doesn’t censor CloudFlare links. I’m not sure if that’s your concern. Search engines with a linear output have to decide on the sequence to list links. The topmost links are the first seen by the user, so it’s important to rank the sites by usefulness. You cannot add a link to the top of the page without pushing another link off the screen. CloudFlare links are poorest quality of results with lower usability because of blockades, CAPTCHAs, and other shenanigans. They don’t get censored by Searxes but they do get low positions to ensure that the more useful links appear first. That’s not censorship; it’s simply smart ranking.
Regarding privacy abuse, I can only guess that you are talking about the row “feeds other privacy abusers”. That means the search engine benefits a privacy abuser, so I see no connection to censorship in that context. E.g. DDG partners with Yahoo to deliver ads to users. That generates ad revenue for Yahoo. Yahoo has a long history of privacy abuse and DDG is helping a corporation that should be boycotted.
Note as well this is the c/privacy community, so the focus here is not censorship but rather privacy.
deleted by creator
Does anyone know where Searxes gets its search data? IIRC DuckDuckGo uses a combination of Bing and its own web crawler, and StartPage uses a variety of providers, including Google.
DDG is a meta-search just as every searx instance is. DDG pays MS Bing and Yahoo (Verizon) for API access. Hence the “feeds other privacy abusers” row.
Startpage pays Google for API access to search results.
Each searx instance does something different. Many scrape DDG/Google/Startpage/Bing etc. for results. Scraping avoids financially supporting those bad actors. But they often get a CAPTCHA, which causes an error that is sent to the user. Some searx instances run their own YaCy crawler. It’s also feasible that a searx instance could buy API access but I don’t suspect this is common.
Searxes does a daily survey of searx nodes that are functioning well, and randomly dispatches queries to an instance. So it’s a meta-search of the decentralized searx network and it has quite good quality control.
