PEPLYST STRUCTURED DATA · 2026.08.12

A structured record
that remembers change.

Each compound keeps its aliases and forms, source-linked claims, evidence boundaries, regulatory status, and dated conclusion history.

ATTENTION AND EVIDENCE

Where attention has moved ahead of human evidence.

The Promise–Proof Gap is not a treatment score. It compares a disclosed attention band with the depth of verified human evidence in Peplyst's reviewed sources.

AttentionAn editorial starting point that will be replaced by first-party behavior after minimum sample thresholds.Human evidenceOnly study-level evidence in people raises this band. Mechanistic and animal findings remain separate.Rate of changeThis signal moves only after a verified change in human evidence.
CompoundAttentionHuman proofPromise–proof gapEvidence velocity
WHAT PEPLYST OWNS

The transformations, not the public inputs.

Method PPG-1.0
01

Identity registry

Aliases, forms, and commonly conflated compounds resolve without erasing material distinctions.

02

Claim ledger

70 propositions preserve identity, use, evidence boundary, conclusion, source, confidence, and effective date.

03

Evidence deltas

ED-0.9 preserves 46 before-and-after records plus explicit conclusion-impact assessments.

04

Precedent graph

RPG-0.6 links 5 controlled comparison clusters while preserving transfer limits.

05

Study registry

PubMed IDs and NCT records normalize into stable study identities with deterministic compound links and an explicit review state.

06

Answer benchmark

Version PB-0.3 uses 140 source-linked questions to test identity, evidence, approval, advisory, and final-action distinctions.

07

Publication memory

Source approval, revision drafting, publication approval, and the public version write are separate attributable records.

08

Citation layer

Stable source IDs, snapshot hashes, derived-record edges, and bibliography formats make provenance portable without copying source bodies.

09

Reliability register

Vendor-neutral model outputs normalize into versioned, reproducible benchmark runs with public aggregate reporting.