Transparency · Source registry

Every outlet we read, and who owns it.

The engine's entire source list, in the open — each outlet's bias bucket, its ownership, and its feed. Our buckets are editorial assessments informed by independent public ratings; cross-check any outlet yourself with the links below.

Loading the registry…

The beat registry — specialist depth, kept separate on purpose

Our desks (National Security, Elections) draw on specialist outlets — defense trade press, regional experts, OSINT researchers — that don't belong in the spectrum registry above: most have no meaningful left–right rating, and mixing them in would distort the bias arithmetic that powers every story's source-spread bar. So they live here, in a second tier: used for desk depth and tips, never counted in the spectrum math, each listed with its ownership and the caveats you deserve to know (including government funding where it exists). The raw list is public JSON at engine/outlets_beat.json.

How the buckets are set — and how to check us

We assign each outlet to one of five buckets: Left, Lean left, Center, Lean right, Right. These are publication-level assessments of the outlet as a whole — story selection, wording, framing patterns — not judgments of any single article. They are informed by the published ratings of three independent organizations that use different methodologies: AllSides (blind bias surveys and editorial review), Ad Fontes Media (multi-analyst article scoring), and Media Bias/Fact Check. If our bucket for an outlet disagrees with your read, check those three — and tell us: [email protected]. Ownership categories are hand-coded from public records and updated when ownership changes. The bucket assignments are the single subjective input to the engine; everything downstream — the source-spread bars, blindspot flags, and coverage-gap callouts on every story — is arithmetic over this list. The registry file itself is public JSON, and the full editorial ruleset is at How the engine works.