Methodology

This page explains what the site does, how each claim is grounded in primary sources, and where its limits lie.

Primary-source, not interpretive

This site reports source text, citations, effective dates, and per-company applicability metadata. It does not interpret regulations, weigh how they apply to your facts, or offer an opinion. Applicability results are threshold-derived statements, “revenue is at or above $X”, with the reasons shown, never advice about what to do.

Citation integrity

Every regulatory claim is meant to pin to an exact span of a stored source snapshot, tied to a version and a retrieval date. Until a claim is backed by a real ingested snapshot it is marked ungrounded seed data and must not be relied on as fact. The current corpus is illustrative seed data, flagged as such on every obligation, so grounded and ungrounded claims are always visually distinct.

Point-in-time history

Regulations change, and our knowledge of them is corrected over time. The corpus preserves two independent dates for every fact: when it was true in the world (valid time) and when we recorded it (transaction time). That is why the as-of-date slider can answer both “what was in effect on a date” and “what did we believe was in effect, as of our knowledge on another date.” Records are never edited in place; a correction is a new record.

Authoritative sources

The intent is to ingest from official sources and APIs (the Federal Register, EUR-Lex, SEC EDGAR, the CARB docket) with provenance, rather than scraping brittle HTML. Changes are detected by content hash so that costly, meaning-aware diffing runs only when a source actually changes.

Where to verify

Always confirm against the cited primary source, and consult qualified counsel for anything that turns on how a rule applies to you. See the status states reference for what each lifecycle state does and does not imply.