clstr.news

CLSTR gives an agent situational awareness of the news: multi-sourced, deduplicated synthesis instead of a stream of near-duplicate headlines. The pipeline reads over 100,000 articles a day from more than 40,000 sources in 170+ countries. Coverage of the same event is gathered across those sources and summarized before it surfaces, so what you get is the shape of the reporting, with a count of how many outlets are covering it and a significance score, rather than one outlet's headline in isolation. CLSTR reports what sources are saying. It does not check whether they are right. Two terms run through every tool: - Cluster: one real-world news event, assembled from many outlets covering the same happening and deduplicated into a single item (one event, not 20 versions of one headline). Each carries a source count (how many outlets are behind it) and a significance score generated by an AI model (1 to 10, where 8 and above is exceptional). - Situation (story arc): an ongoing storyline that groups related clusters over time. It keeps a maintained summary and a full chronological timeline, so you see not just what happened but how it developed. This is the memory: you can answer "what has happened here over the last six weeks", not just "what is the latest". Which tool to use: - search_situations: find coverage by topic, entity, company, country, or event. - get_top_situations: what is developing right now, optionally within a category and a time window (24h, 7d, 30d). - get_situation_timeline: the whole history of one situation, in order, to brief on how it developed. - get_cluster: the detail of one specific event, with its underlying source articles. Every result carries canonical clstr.news links. Always cite those URLs so a reader can check the sources.