VERIFICATION POLICY

Sources and verification.

Ramen AI Wiki is an archive of 'who said what, how' — not an archive of facts. The following formalises our sourcing, confidence, and privacy policies.

confirmed / needs_source / disputed

Each claim, relationship, person, and timeline event carries a confidence_status. confirmed = widely told; needs_source = awaiting source; disputed = multiple viewpoints exist.

Attributed viewpoint.

Whenever possible, claims carry a speaker (who) and reliability (first-person / interview / community / disputed). Assertive phrasing such as 'fact', 'truth', or 'absolutely' is not used.

Source URLs — verification only.

Source URLs are kept internally for verification and shown subtly (dotted-underline). We do not embed giant players or autoplay video. Google-derived data is marked attribution-required.

We do not republish transcripts.

Full YouTube / interview transcripts are never republished. Only summaries, quotes, and claims are stored. Bulk transcript paste is blocked at submission time via heuristic.

Disputed is shown, not hidden.

Disputed relationships are drawn as dashed red edges with a disputed badge. We do not silence one view to keep the page tidy — we publish that multiple views coexist.

Privacy.

We only handle publicly stated, officially published, or media-covered information. Private life, family, addresses, or phone numbers are never stored at the schema level.