News Tracker: Real-time system tracking how news stories spread across 200k websites

What it does

News Tracker monitors approximately 200,000 news RSS feeds and clusters related articles in near real-time. It groups stories that are substantively the same, showing which outlet published first, how quickly the story spread to other publications, and how the narrative changed as more media outlets covered it. The system uses Snowflake's Arctic model for embeddings and HNSW for similarity searching to identify related articles across the feed network.

Who it is for

The service targets people who want to understand news propagation patterns. This includes journalists tracking how stories develop, media analysts studying narrative evolution, researchers examining information spread, and news enthusiasts wanting to see breaking stories and trending topics across thousands of sources simultaneously.

Pricing

The site does not show prices.

How it stands out

News Tracker focuses specifically on story propagation rather than just aggregation. Instead of simply listing articles, it visualizes the timeline and path a story takes across the media landscape. By showing the originating outlet, propagation speed, and narrative shifts, it adds analytical layers that basic news aggregators do not provide. The system's ability to handle approximately 200,000 feeds in near real-time suggests substantial infrastructure investment.

What a founder should check

A rival builder should investigate three key areas. First, examine the incumbent competition carefully. Services like Google News, Mediacloud, and Factba.se already aggregate news at scale and track story spread to some degree. Understanding their feature gaps and user frustration points is essential. Second, assess switching costs for users currently relying on existing tools. If publications or researchers have built workflows around current aggregators, migration friction may limit growth. Third, validate whether the technical moat holds over time. The current implementation relies on openly available embeddings models and standard similarity search algorithms, meaning competitors could replicate the architecture relatively quickly if market demand proved real. The real moat would come from data network effects—access to feeds others cannot get—or from proprietary clustering accuracy that meaningfully exceeds alternatives.

Thinking of building something like this?

Every launch here is a competitor to somebody's idea. If yours is close, check it against the market before you build: the Full Check names the rivals, the prices and the gaps.