UPDATE: Until further notice, my Searx instance is offline. I suggest using the primary instance of Searx, at Searx.me.
I’ve setup my own private instance of Searx, located at blackholeroutercondition.com.
Searx is described by its author as a “privacy-respecting, hackable metasearch engine“. Because of the resources required to index web pages (i.e., CPU and disk), and the fact that the web is constantly changing, most individuals/businesses can’t afford to setup their own search engines. Because of this, only a handful of search engines exist that actually index their own pages (e.g., Google, Yahoo, Baidu, Yandex, etc…). A metasearch engine (e.g., Ixquick, Startpage, Dogpile) aggregates results from multiple sources into one place, as shown below.
Searx, however, takes metasearching a step further. Not only is Searx open source, but it removes tracking, so companies (e.g., Google) won’t be building a profile about you. Searx also includes a bunch of other features, including Opensearch support (so you can set Searx as the default search engine), the ability to configure search engines and categories (via a cookie), multiple languages, DuckDuckGo-like !bang functionality, and you can self-host it!
If you’d like, you can stand up your own instance of Searx using the instructions on GitHub. The primary instance of Searx is located at searx.me, and the community keeps a running list of public instances on GitHub as well.
My instance of Searx is completely stock, except that I’ve tweaked it to lean more towards Google/Bing/Wikipedia, rather than Yahoo or DuckDuckGo. I’ve also turned off logging in Nginx, so you can search truly anonymously. I do use Piwik to track how many users visit my Searx instance. Since it’s the same instance of Piwik I’m using to host loganmarchione.com, the same rules apply.