How we verify
The status ladder lives here. Not on the homepage.
We publish one kind of thing: plain-language explanations of published GMP standards, so an operator can understand what the text requires and then go read the text itself. Three rules govern everything.
Expert-signed pages
0
Empty is the correct state. Automation may only lower trust. It may never raise it.
- 01DraftUnverified. Readable as education. Not rendered as a signed fact. Not submitted for indexing.
- 02CitedA primary source is attached. A person has not yet signed that the page says what the source says.
- 03Expert-signedA named, credentialed reviewer confirmed the page against that version of the source.
1. Every regulatory claim traces to a named primary source
We cite the publisher of record and nothing lower: the European Commission’s EudraLex Volume 4 (including Annex 7), the EMA for the GACP guideline, the ICH database for ICH Q7, the eCFR for 21 CFR Parts 210 and 211, and the FDA/DEA primary record for US federal questions. Industry bodies, consultancies, and vendor blogs are used to locate a source, never as the source. A regulatory value with no primary citation does not get published — the page renders the general framing and withholds the figure.
2. Every claim carries a status on a three-rung ladder
A regulatory value is one of three things. DRAFT: unverified, not rendered as fact. SOURCE_LINKED: a primary source is attached. SME_VERIFIED: a credentialed reviewer opened the source and confirmed it says what we claim. This rung is human-only — no automation can grant it. Sourced is not the same as verified. Only SME_VERIFIED takes the draft label off a page. Automation on this site can lower a page’s trust state. It can never raise it.
3. A credentialed QP or RA professional reviews a guide before it is indexed
Review is separate from sourcing. Sourcing proves a value exists in a document. Review proves a named person with the right credential agrees the page is a fair account of that document. No guide carries an expert signature yet — and no page pretends otherwise. We do not invent reviewers, dates, or certificates to make a page look finished.
See also editorial and ethics and the regulatory timeline.