Search the headlines of 325 million news articles back to October 2019. Data is primarily sourced from GDELT. Beginning in September 2026, data is also sourced from blogs and news feeds.
The GDELT project extracts geolocated event data from millions of news articles over time and makes the resulting data available for download. Beginning in 2019, this data includes the headline (but not the body) of each article. These headlines alone are a rich and entertaining source of information about global attention, and this website makes it possible to quickly search these headlines for words or strings, then chart the result. You can filter articles by source or date range, compare term searches to each other, and share direct links to interesting findings. You can also read the articles (as long as the original source’s link still lives) by following the result links.
To expand coverage of news sources not present in or no longer collected by GDELT, we also collect headlines from first-party news feeds, Substack, and Medium. Data from these non-GDELT sources starts from September 2026; by default these additional sources are not included in search results.
Please also see the source code for this project, which was “vibecoded”.
Other projects:
Contact: loft4map [at] gmail.com
Try: ceasefire, climate change, zelensk (substring), election on bbc.com, everything from irishtimes.com, tariff, oldest first; or compare ukraine, gaza and sudan month by month.
Count Scale
A headline counts as one “story” on its first appearance in any source in one week. A headline counts for one “outlet” the first time it appears at a given source in one week. A headline counts for one “article” every time it appears whatsoever, regardless of previous appearances globally or for a given source.