The Front Page

Start with what it refuses to do. There are no accounts, so there is nothing to personalise against. No recommendations, so no filter bubble. No ads, sponsored placements, or affiliate links, so nothing competes with the ranking. No engagement tracking, no infinite scroll, no notifications.

What is left is a front page: the day’s headlines from more than eighty countries, the way a wire editor would have laid them out.

How a story earns its place

Every two hours a pipeline pulls from over a thousand RSS feeds, scores each headline by journalistic importance, clusters stories that several outlets are covering, translates non-English headlines, and writes the whole site out as static HTML.

The ranking signal that matters most is corroboration. A story carried by five independent outlets in a country outranks one carried by a single outlet, because the question the page is trying to answer is not what will hold your attention but what did the press of that country think was important today. Each headline names its sources, so you can see the basis for the ranking and disagree with it.

Drilling in

Each of the eighty-odd countries has its own page: up to twenty headlines, grouped into story clusters, with sources listed. Countries are grouped by region, and the site installs as a progressive web app if you want it on a phone home screen.

It deliberately stops short of the article. Headlines and links only — no scraped article text, no archive, no full-text search. The site is an index of what is being reported, not a replacement for the people reporting it.

Cost

The whole thing runs on free infrastructure: GitHub Actions for the pipeline, GitHub Pages for hosting and the CDN. Total ongoing cost is the domain name. That constraint is the reason the architecture is what it is — a static site regenerated on a schedule has no servers to pay for, no database to run, and no way to degrade under traffic.