How Zen Reader works

Most news apps are built to keep you looking. Headlines are written to make you anxious or curious enough to tap, the same story appears ten times from ten sites, and the feed never ends, so there is always one more thing to check. It is easy to close the app feeling worse and somehow still not caught up.

Zen Reader is the opposite. You choose the topics you care about. It reads the news on them for you, puts each story in one place, writes it plainly, and gives you a short page that ends. No trending lists, no counts of likes or followers, no streaks, nothing designed to pull you back.

Every few hours, for each topic, Zen Reader looks around the web the way you might if you had the time: news searches, the sections of newspapers and magazines that cover the subject, blogs and the topic's own feeds, and conversations on Hacker News and Reddit. It checks when each article was really published, so old news doesn't pass for new, and it leaves out sites that only copy other outlets' work.

A big story is covered by dozens of sites at once, each with a different headline. An AI model reads the headlines and the first lines of every article and notices what they have in common: the same people, places, numbers and dates. Articles about the same event are grouped into one story, and different events stay apart, even when they are about similar things.

Each story leads with where the news came from: the organization that announced it, or the outlet that reported it first. The other reports are listed underneath, so you can read more than one side if you want to, without seeing the same story over and over.

Every story gets a new headline and a sentence or two saying what happened, written to inform rather than to provoke: no shouting, no cliffhangers, no “you won't believe”. Nothing is added that isn't in the articles themselves, and when a story rests on unnamed sources or a leaked document, it says so. The original articles are always one tap away.

Home shows the newest stories from all your topics, a couple from each, and each topic has its own page with the rest. Stories older than a couple of days settle lower. At the bottom there is a small drawing, and that is all: you have read what matters, and you can put it down. If you like, a short email can bring the same thing to you once a day or once a week.

Each topic's stories are written once and shared by everyone who follows it, so the AI model reads the news, not you: it never sees what you read. The model can get things wrong now and then, which is why every story links to the original reporting. What is stored about you is on the privacy page.