We don't tell you what to think. We show you how it's being told.

The problem we're solving

Trust in media is at a historic low. Not because outlets are lying, but because they frame. The same set of facts can become completely different stories depending on which details are emphasised, which voices are quoted, and what angle is taken.

Most people sense this. But seeing it for yourself meant opening ten tabs and doing the comparison manually. Nobody has time for that. So most people just pick an outlet they already agree with and stay there.

Clarrus does the comparison for you. Every story. Every day. Automatically.

Same story. Same day. Three completely different narratives.

How it works

1. 1 We collect. Clarrus continuously ingests articles from 65+ established news outlets - left, centre, and right-leaning sources from the UK, US, and internationally.
2. 2 We cluster. Articles covering the same underlying event are grouped into a single "story" using semantic similarity, shared entities, and timing.
3. 3 We analyse. For each story, an AI model reads coverage from each political lane and identifies: what facts are universally reported, where emphasis diverges, and how the narrative differs.
4. 4 You decide. The analysis is presented neutrally - you see the framing differences clearly, without being told what to make of them.

Stories are re-analysed as new coverage appears. What you read today reflects the current state of coverage, not a snapshot from when the story broke.

What Clarrus is not

• — Not a news publisher. We don't write articles, break stories, or have editorial opinions. Every analysis is generated from third-party journalism.
• — Not a fact-checker. We analyse how things are reported, not whether they're true. We don't verify claims made by individual outlets.
• — Not politically aligned. Clarrus doesn't take sides. Our analysis treats every outlet's coverage as data, not as right or wrong.
• — Not infallible. AI analysis can contain errors or miss nuance. We work to improve accuracy continuously, but this is a tool to aid understanding, not replace it.

How Clarrus differs from similar tools

Some media bias tools show you which outlets covered a story — and whether left or right outlets were more likely to report on it. That's useful for spotting coverage gaps: a story only one side picked up.

Clarrus goes one step further. When the same story is covered everywhere, we show you how the framing differs. The same arrest, the same vote, the same speech — read by a left-leaning outlet versus a right-leaning one — can tell completely different stories without a single false fact. That's the analysis Clarrus is built around.

Most tools tell you who's reporting. Clarrus tells you what they're emphasising, whose voices they're quoting, and what angle they're taking — across every lane, on the same event, side by side.

How we classify outlets

Each source is assigned a broad political orientation (left, centre, or right) based on long-term editorial patterns across many stories. Classifications are reviewed periodically and are not based on any single article or topic. They reflect observed tendencies in framing, not endorsement or condemnation of any outlet.

We cover a deliberate mix: outlets that readers across the political spectrum would recognise as authoritative within their tradition. Our goal is breadth and balance, not a perfect 50/50 split.