# CLSTR

CLSTR is a situation monitoring platform that groups related news articles into Clusters and tracks evolving storylines as Situations.

## Concepts

- Cluster: a group of articles from different sources covering the same discrete news event.
- Situation: a longer-running story arc made from related Clusters over time.
- Topic: a user-saved semantic search.
- Monitor: an alert rule that notifies users when a Topic matches a new Cluster or a Situation materially changes.
- Briefing: a scheduled summary of what changed since the previous briefing.

## Important pages

- Home: https://clstr.news/ (Situations feed)
- All situations index: https://clstr.news/situations/browse (every public situation, paginated)
- Concluded situations: https://clstr.news/situations/archive
- Clusters: https://clstr.news/clusters
- Situation pages: https://clstr.news/situations/<slug>
- Cluster pages: https://clstr.news/cluster/<slug>
- Rooms (curated collections around one major event): https://clstr.news/rooms
- About: https://clstr.news/about
- Contact: https://clstr.news/contact

## Feeds

- Situations (Atom): https://clstr.news/feed/situations.xml
- Clusters (Atom): https://clstr.news/feed/clusters.xml
- Clusters by category (Atom): https://clstr.news/feed/clusters/<category>.xml

## Machine-readable access

- Developers/API (keyed JSON, situations, clusters and semantic search): https://clstr.news/developers
- API reference (markdown): https://clstr.news/developers.md
- Developer recipes (morning brief, dashboard poll, topic monitor): https://clstr.news/developers/recipes
- Which news API an agent should use: https://clstr.news/news-api-for-ai-agents
- MCP server for news (connector setup for Claude, ChatGPT and frameworks): https://clstr.news/mcp-server-for-news
- The REST API is keyed and versioned (v1). Create a free key from https://clstr.news/developers, then send it as an Authorization: Bearer header.
- MCP server (JSON-RPC over HTTP, the recommended path for an agent): https://mcp.clstr.news/mcp. Four read-only tools: search_situations, get_top_situations, get_situation_timeline and get_cluster. Some of them answer with NO key at all, capped per network, so a client can be pointed at it before signing up. Search and get_cluster always need a key.
- Markdown mirrors: append .md to any situation URL, e.g. https://clstr.news/situations/<slug>.md (also advertised via link rel="alternate" type="text/markdown" on each situation page)
- Sitemaps: https://clstr.news/sitemap.xml, https://clstr.news/sitemap-situations.xml, https://clstr.news/sitemap-news.xml

## Notes

CLSTR does not publish original reporting. It groups and summarizes coverage from linked source articles. Source links are shown on Cluster pages.
CLSTR allows AI systems to read and cite this site. Please attribute and link to the canonical clstr.news URL.
Training on this content is not permitted (Content-Signal: ai-train=no).
Pages update as situations develop: coverage follows the original reporting rather than breaking it, and gains sources and detail over time.