AI-curated global news that evaluates article relevance to humanity across four issue areas, with transparent methodology, a free public API, and no ads or tracking.

Most news feeds optimize for clicks, not importance. The result is a stream skewed toward outrage, celebrity, and conflict — while stories  that actually matter to humanity get buried. Actually Relevant flips that. It crawls >80 sources across 5 languages and every continent, then uses LLM analysis to evaluate each article's relevance to humanity across four issue areas: Human Development, Planet & Climate, Existential Threats, and Science & Technology. Only the most relevant stories make the cut. Each one on the site is more relevant to humanity than 98% of stories on traditional newspaper front pages. The full methodology is published at actuallyrelevant.news/methodology. A positivity slider lets readers counteract negativity bias directly: adjust it to surface more solution-oriented and constructive coverage without filtering out hard news entirely. Other features include RSS feeds, a newsletter, widgets, and a free public API (no auth required, OpenAPI documented at /developers). It's a solo side project built in Berlin, Germany, with a social entrepreneurship mindset: non-commercial, no ads, no tracking. The project is actively seeking a long-term institutional owner in journalism, civic tech, or effective altruism who can take it further. A /stewardship page on the site explains the vision. People can help by sharing the site with anyone interested in global news curation, connecting us with potential steward organizations, or using the free API to build on top of it.