Interview: Searx, the web search engine that respects your privacy

Search engines are a fundamental part of the internet, but most people only use one. Google has a market share of 86 percent worldwide and 95 percent in Germany. There are alternatives, some also promise data protection, but most are commercial black boxes. But there is another way, like Searx – a privacy-friendly, decentralized and non-commercial search engine. An interview with Adam Tauber, free software developer and Searx maintainer from Hungary.

This interview is also available in German.

netzpolitik.org: What is Searx?

Adam Tauber: Searx is a free meta-search engine that respects privacy. Users submit search queries to Searx and Searx sends the queries to over 80 other search engines, gathers the results and ranks them.

Privacy-respecting means Searx does not track your moves and doesn’t hand over or sell your personal data to third parties. Only absolutely necessary information is forwarded to other search engines, like your search query and language.

Try it at Searx.me.

netzpolitik.org: What makes Searx different from other privacy-focused (meta)search engines like DuckDuckGo or Startpage?

Tauber: There are two main differences: Searx is free software and decentralized.

Free software means all source code is public. Anyone can verify whether Searx does something shady or not. Anyone can use it for any purpose and make changes.

DuckDuckGo and Startpage are both closed source software. Users have no way to verify if they really let you search without being tracked. You have to trust them blindly.

Searx is decentralized, anyone can take the source code and run their own instance. There are numerous public Searx instances. If you can’t or don’t want to run a server yourself, all you need to do is find a Searx instance you like or trust.

DuckDuckGo and Startpage are both centralized, and run by for-profit companies.

netzpolitik.org: And MetaGer, another meta-search engine focused on privacy, based in Germany?

Tauber: Unlike the others, MetaGer is also free software, like Searx. You can host it yourself, so you don’t have to trust third-parties. Maybe MetaGer is more appealing to a German audience. But there are some differences between MetaGer and Searx.

The main goal of Searx is to provide a search engine with an emphasis on privacy protection. MetaGer stores search results in a database, Searx runs without any database. Searx does not store any past queries.

Searx is more customizable. Users can change and adjust their theme, search settings and privacy settings. The user interface is available in 30 languages, while MetaGer has three.

A unique feature of Searx is its support for scientific searches. Searx has interfaces to multiple search engines, aggregating scientific papers from repositories like BASE, Google Scholar or PubMed. To foster a free flow of information and knowledge, Searx has a plugin to find open access versions of academic papers.

netzpolitik.org: Are Searx search results as good as the ones from Google?

Tauber: Searx is only as good as its configuration. We provide sane and safe defaults. But to get the results you need, you might want to enable or disable certain search engines and maybe fine-tune timeouts. Correctly used, search results can be as good or even better.

Searx is superior to Google in regard to privacy and filter bubbles. If you only enable the Google engine in Searx, you can use Google without exposing unnecessary personal information. Searx results are unbiased and non-individualised, since it does not know more than necessary about users.

netzpolitik.org: What’s wrong with Google?

Tauber: Google is a huge international company which wants to make money wherever it can, including by exploiting your personal data. There are many articles detailing what is wrong with Google.

netzpolitik.org: How many people use Searx?

Tauber: I don’t know, there are many Searx instances beyond my control.

I can only estimate the number of search queries on the instance I run: Searx.me. This web server is restarted after processing one million requests, this is the only information we are logging. Using this information we can estimate that Searx.me has around five to ten million search requests per month, that’s two to four every second.

netzpolitik.org: Who builds and runs Searx?

Tauber: The core team consists of three maintainers. But we have many other developers who contribute significantly. For example, Johannes Schauer maintains a package for Debian-based operating systems.

Public Searx instances are run by individuals and organizations who care about privacy. There’s a list of public instances in our wiki.

netzpolitik.org: How can people support Searx?

Tauber: Anyone can contribute on Github – by opening an issue, coming up with new ideas or by submitting contributions.

Furthermore, people can run public instances. This makes Searx available to people who do not want to run their own servers.

If someone wants to support the development with money, there are a few options to donate.