Developer Lyra Rebane has launched Xikipedia, a web app that displays Wikipedia entries in a social media-style feed. The app draws content from Simple Wikipedia and uses a recommendation algorithm that operates entirely locally, without collecting or sharing user data.
The Xikipedia landing page states: “It is made as a demonstration of how even a basic non‑[machine learning] algorithm with no data from other users can quickly learn what you engage with to suggest you more similar content.” It adds: “No data is collected or shared here, the algorithm runs locally and the data disappears once you refresh or close the tab.”
i made a version of wikipedia you can doomscrollxikipedia.org
Users can filter the feed to show entries from specific categories, including custom ones. Each post presents a summary of a Simple Wikipedia article. Liking a post increases the chances of seeing future posts from the same category, its parent categories, and articles linked within it, according to Rebane.
Clicking or tapping a post opens the full Simple Wikipedia article. The app selects articles at random, which means extended scrolling may expose users to not-safe-for-work (NSFW) material. Xikipedia requires loading about 40 MB of data, resulting in a short initial wait.
Simple English Wikipedia holds more than 278,000 articles, offering hundreds of thousands of potential posts. However, updates occur less frequently than on the main Wikipedia site. For instance, one musician’s article lacked entries for their two most recent albums in the discography section.
Xikipedia functions like the former StumbleUpon service, providing a way to discover new topics through random yet personalized article suggestions as a break from traditional social media feeds.
