searx is a freemetasearch engine, available under the GNU Affero General Public License version 3, with the aim of protecting the privacy of its users. To this end, searx does not share users' IP addresses or search history with the search engines from which it gathers results. Tracking cookies served by the search engines are blocked, preventing user-profiling-based results modification. By default, searx queries are submitted via HTTP POST, to prevent users' query keywords from appearing in webserver logs. searx was inspired by the Seeks project, though it does not implement Seeks' peer-to-peer user-sourced results ranking. Each search result is given as a direct link to the respective site, rather than a tracked redirect link as used by Google. In addition, when available, these direct links are accompanied by "cached" and/or "proxied" links that allow viewing results pages without actually visiting the sites in question. The "cached" links point to saved versions of a page on archive.org, while the "proxied" links allow viewing the current live page via a searx-based web proxy. In addition to the general search, the engine also features tabs to search within specific domains: files, images, IT, maps, music, news, science, social media, and videos. There are many public user-run searx instances, some of which are available as Tor hidden services. "Meta-searx" sites query a different random instance with each search. A public API is available for searx, as well as Firefox search provider plugins, and an Android app.
Across all categories, searx can fetch search results from about 82 different engines. This includes major search engines and site-specific searches like Bing, Google, Reddit, Wikipedia, Yahoo, and Yandex. The engines used for each search category can be set via a "preferences" interface, and these settings will be saved in a cookie in the user's browser, rather than on the server side, since for privacy reasons, searx does not implement a user login model. Other settings such as the search interface language and the search results language can be set the same way. In addition to the preferences cookie, it is possible on each query to modify the engines used, search categories selected, and/or languages to search in by specifying one or more of the following textual search operators before the search keywords.
!category — Search the specified category instead of the default ones.
?category — Search the specified category in addition to the default ones.
!engine — Search the specified engine instead of the default ones.
?engine — Search the specified engine in addition to the default ones.
:language — Search for results in the specified language instead of the default one.
The ! and ? operators can be specified more than once to select multiple categories or engines, for example !google !deviantart ?images :japanese cow.
Instances
Any user may run their own instance of searx, which can be done to maximize privacy, to avoid congestion on public instances, to preserve customized settings even if browser cookies are cleared, to allow auditing of the source code being run, etc. Users may include their searx instances on the editable list of all public instances, or keep them private. It is also possible to add custom search engines to a self-hosted instance that are not available on the public instances. Another reason to use different searx instances, and/or to run one's own, is that as of 2019, Google has begun to block some instances, including some of the IP addresses used by searx.me, from querying it, resulting in a "google " error. In response, some instances have been modified to silently skip trying to search with Google, even when it's the only engine specified.
