Methodology

Source grading

Every story is traced to its publishers. We prefer primary sources (official statements, government releases, on-the-record interviews) and established regional outlets. Aggregated stories list each contributing source so readers can verify the chain of reporting themselves.

Verdict scale (fact-checks)

• Verified (5/5) — corroborated by multiple independent reliable sources.
• Likely True (4/5) — strong evidence, one independent source pending.
• Mixed / Disputed (3/5) — partial accuracy or contested by credible sources.
• Misleading (2/5) — technically accurate framing that omits material context.
• Likely False (2/5) — evidence points against, no credible corroboration.
• False (1/5) — clearly contradicted by reliable evidence.
• Unverified / FAKE NEWS (1/5) — no credible source could be located.

Corrections policy

When a story is materially updated after publication, we surface an "Updated" note tied to the article's modified date. Substantive corrections are noted at the foot of the article. Reach the editorial desk atour contact page to flag errors.

AI and editorial oversight

We use automated pipelines for translation, source clustering, and a first-pass rewrite. Every claim is cross-checked against the original source by an AI faithfulness gate, and items failing the gate are held for human review before publication.