About brokenews

Every story is told differently depending on who's telling it. brokenews puts those versions next to each other, left, center, and right, using each outlet's own reporting, and shows what one side covered that another left out.

We don't tell you who's right. We show you the spread, with the receipts, so you can read the framing yourself.

This is a media-literacy project as much as a publication. The point isn't to make you trust us. It's to make you read across, notice the framing, and trust any single feed a little less.

How much of this is machine made

Most of it, and here is which parts.

Machines do the work: fetching the feeds, grouping articles into stories, pulling out claims and quotes, comparing framing across the spectrum, drafting the neutral account, writing the social copy, and reading the audio.

A person decides the frame: which outlets are tracked and where they sit on the spectrum, who owns them, what counts as a correction, and which stories go out on social. A person can take any story down, and does.

Stories reach this site automatically once they clear the automated checks. We are telling you that plainly rather than implying a person reads each one first. What a person guarantees is the rules the machine follows, and being answerable when it gets one wrong.

Nothing on this site is written by a journalist we invented. The byline is the organisation, because the organisation is what is accountable.

When we get it wrong

We will. Corrections are public and dated, on the corrections page. If a fact, a quote, a source, or a spectrum label is wrong, tell us and it gets fixed in the open rather than quietly edited.

Reuse the analysis

We license our analysis data to you; CC BY 4.0 is the governing deed. Analysis data is licensed CC BY 4.0. Credit brokenews and link to the story page you took it from. Source headlines, story titles taken from a source headline, and quoted passages remain the property of their publishers and are quoted for comparison, not licensed onward.