About
The Dutch news day, filtered.
De Krantenkoppen reads every Dutch newsroom so that you do not have to read all of them. It gathers each day's reporting story by story, writes every story up in English, and then sends you back to the papers that did the work.
Who this is for
About forty newsrooms cover this country. On any given day a dozen of them file the same story, none of them file all of yours, and the whole thing arrives as more than a thousand headlines a day in no particular order. Following that properly is a job.
This is for anyone who wants the country without the job. You live here and your Dutch stops somewhere short of the eight o'clock news. Or your Dutch is fine and your patience for forty feeds is not. Or you have left, or have not arrived yet, and you would still like to know what the Netherlands is arguing about this week.
The reporting exists. It is simply spread across dozens of newsrooms, mostly in Dutch, and none of it slows down for you. This site reads all of it every ten minutes and tells you what happened.
How a story gets onto this page
- 01
We read the feeds, not the sites
Every ten minutes we fetch the public RSS feeds that Dutch newsrooms publish for exactly this purpose — the national broadcaster, the quality dailies, the financial press, thirteen regional broadcasters, the city outlets and the English-language titles. Nothing is scraped. Nothing behind a paywall is opened.
- 02
We group the coverage into stories
Each article is compared against the ones already in hand and joins the story it belongs to, or starts a new one. Twelve papers on one court ruling become one entry here, not twelve, and the number of newsrooms that filed becomes evidence of how much the story mattered.
- 03
We write our own account
A language model reads the whole group and writes a title and summary. Both are ours. Neither is a translation of anybody's headline, and neither is anybody's copy.
- 04
We rank by the day
How recent the story is, how many independent newsrooms filed on it, how fast that coverage is still growing, and which subject it falls under. There is no personalisation and no editor with a favourite. Everyone who opens the site sees the same front page.
- 05
We hand you back to the source
Every story names the outlets that covered it and links straight through to their reporting.
What we never publish
A publisher's headline. Not translated, not adapted, not “inspired by”.
A publisher's text. No standfirsts, no opening paragraphs, no lifted sentences. Our summary describes a story. It is not a substitute for reading one.
A publisher's photographs. We hold no images from anyone's feed on any page of this site.
Anything between you and the source. No advertisement beside someone else's reporting, no interstitial, no sign-in wall, no “continue reading here”.
Where the machine is, and where it is not
The titles and summaries here are written by a language model from the source articles. Grouping the coverage and translating it are machine work too.
What is not machine work: which outlets are read, how much weight each subject carries, and the rule that nothing verbatim is ever republished. Those are decisions, made by people, and written into the code rather than left to a model's judgement.
A model can still misread a story — compress two facts into one, drop the qualifier that mattered. When our summary and the linked reporting disagree, the reporting is right.