OSMS

The problem: The same news reaches you many times: three tech sites, a newspaper in another language, a podcast, a research paper, each as a separate item, and today's AI readers just add ten separate summaries. You reread the same event, can't tell what's actually new since you last looked, miss where sources disagree or were corrected, and reports in another language never meet. What OSMS does: OSMS (One Story, Many Sources) reads every source you follow (articles, arXiv papers, podcasts, social posts) in any language, and groups the reports of one event into one story. Each story has one summary in which every sentence names its sources and links to the original (OSMS links, it never copies). It shows what's new since your last visit, where sources disagree, side by side, and a dated timeline that keeps corrections. Quotes stay in their own language with a labelled translation, so a Danish and an English report land in one story. "How OSMS decided" shows every model call behind a story. Each morning OSMS makes a short podcast of your top stories in which every sentence cites a checked fact, and you can ask a story questions that are answered only from its checked facts. Add any site, feed or social handle by pasting it; saved stories are the same on the web and the Mac. The result: One story per event instead of ten tabs: you read only what changed, and you can check every claim against its source.