For pharma copy under MLR review

1

1 · Wilding et al., N Engl J Med 2021 · p.989 ¶3 s1 · 96% · Approved match

finds the claims, then the evidence.

Draft the way you always have. CRx picks out every statement that needs a reference, pulls the passages that support it, page and paragraph included, and lines up the wording the brand has already had approved. The hunting is done before review starts.

HCP · Detail aid · Draft

Weight Management Detail Aid

142 words · 1 library block

Weight loss

In adults with obesity and without diabetes, once-weekly semaglutide 2.4 mg produced a mean weight loss of 14.9% at 68 weeks versus 2.4% with placebo. More than four in five participants lost at least 5% of their body weight, and half lost 15% or more.

Cardiovascular outcomes

In adults with established cardiovascular disease and overweight or obesity, semaglutide cut the risk of heart attack, stroke or cardiovascular death by 20% versus placebo, an effect that emerged within the first months of treatment.

Approved matchNear match · verifyRewrite suggestedNeeds evidence

Locations are computed from the parsed source, not guessed by the model. Confidence comes with the reasoning a reviewer can check.

How it works
01

The library reads itself.

Upload the trials, prescribing information and guidelines. Each page is parsed, cross-checked against the PDF's own text layer, read again by a vision model where the two disagree, described figure by figure, and indexed by meaning.

02

Every claim gets its evidence.

Check claims finds each referenceable statement in the copy, retrieves the library's best passages, reranks them and grades each against the claim as worded. Exact and near matches to approved claims reuse the approved references; close-but-different wording gets a rewrite.

03

Review, then export.

Comments anchored to the copy, with mentions. Select and order the references you'll cite. Export to Word with superscript citations, a reference list and an annotated appendix for the reviewer.

Evidence, not vibes

A passage that contradicts the claim scores near zero, and says why.

The grader judges the claim exactly as worded: population, endpoint, magnitude, comparator, product. A related passage that supports a weaker version scores as such. Nothing is invented; every excerpt is copied verbatim from the source.

90–100Same fact, same scope. A reviewer would accept it as the reference.
70–89Substantiates the claim; one detail differs or is unstated. Verify before use.
40–69Partial support, or support for a weaker version of the claim.
0–39Topically related, unsupportive, or contradicting.
The claims library

What can the brand say?

Ask the library directly. Discover claims drafts candidate statements from the evidence, each pinned to one passage with its excerpt. Curate them, send them to MLR, and once approved they become the wording every future check matches against.

What can ORVEXA say aboutsemaglutide and cardiovascular risk↵ draft claims
Candidatedrafted from the libraryProposedcurated, ready for MLRApprovedmatched on every check
The galaxy

A map of what the library knows.

Every passage placed by meaning. Papers cluster with the papers they agree with; a search lights the passages that answer it. Near things say similar things.

STEP 1 · weight managementSELECT · cardiovascular outcomesSUSTAIN-6 · type 2 diabetesPrescribing information
Flow State

Built for the people who write the copy.

⌘. hides everything but the paper. Approved content drops in with an @. The evidence is there when you look for it and quiet when you don't.

⌘. Flow State — just you and the paper
Projects / UntitledSaved · v1 · just nowRequest review
BIUSH1H21.@

Start writing. Type @ to insert approved content.

The name

say

x is a cross-reference. Every claim in the copy, cross-referenced to the passage, page and paragraph behind it. That is the whole product.

x is also an Rx. It is written for the people who write pharma copy, and read by the people who review it.

Less hunting. More writing.

Bring the library. CRx reads it, checks the copy against it, and hands review a piece that already shows its evidence.