A search tool I built for a volunteer Chinese science group: it reads RSS from the sites they already trust, ranks stories with embeddings, and helps editors find fresh English articles worth translating for students

ArticleSearch

I help a WeChat-run science outlet that turns English research news into Mandarin for everyday readers. The editors kept refreshing the same handful of homepages by hand, often landing on the same stories, with no quick way to see what was new or overlapping. This app pulls from the RSS feeds they already use (structured, mostly outside paywalls), caches them on the server, and ranks by embedding similarity there when you sort by relevance. I did experiment with base64-packing the vectors and searching in the browser, but server-side caching won out in the end. I added more feeds later as a demo; the point stayed the same: spot interesting science articles before anyone sits down to translate.