KOPENI ONE — Peptide Half-Life Audit Report

Date: 2026-10-22 Scope: Read-only audit of all stored peptide half-life data and every app-wide consumer Status: No records, protocols, doses, logs, or calculations modified. Awaiting approval.

---

Executive Summary

The database contains 38 peptide records with a single half_life_hours field — no unit, source, evidence, species, route, or formulation metadata. The half-life value feeds four separate inline implementations of the same decay model (level = dose × 0.5^(hours/half_life)), plus display surfaces and two AI-prompt contexts.

Findings:

• 3 verified mismatches where the stored value contradicts available pharmacokinetic evidence (GHRP-6, BPC-157, AOD-9604)
• 1 duplicate compound with two separate records (Bremelanotide / PT-141)
• 12 evidence-limited values for research peptides with no human PK data, where the stored value is a community estimate presented without caveat
• 4 minor discrepancies where the value is close but not exact to the authoritative source
• 15 values verified as consistent with available evidence
• 3 values with ambiguous/conflicting evidence needing explicit caveats (MK-677, GHK-Cu, Epitalon)

Structural issues:

• Single numeric field with no metadata (unit, source, species, route, evidence status)
• Decay model duplicated across 4 files instead of a shared authoritative source
• No distinction between elimination vs. terminal half-life, or human vs. animal evidence
• Duplicate compound record risks divergent frequency and metadata

---

Consumer Inventory

Consumers that READ half_life_hours for computation

Consumers that DISPLAY half_life_hours

Consumers that FEED half_life_hours to AI

Coach Integration

base44/functions/kopeniCoach/entry.ts and base44/shared/coachProvider.ts contain no direct half-life references. The Coach does not independently compute clearance or activity from half-life. However, peptide protocols and levels flow into Coach context via the dashboard's context_payload, so half-life-derived level estimates can indirectly inform Coach responses.

---

Complete Audit Table

Tier 1 — FDA-Approved / Clinically Characterized (Human PK Data)

Tier 2 — Research Peptides with Published Human/Animal PK Data

Tier 3 — Research Peptides with No Published PK Data

---

Issue Summary

❌ Critical — Mismatched Values (contradict available evidence)

1. GHRP-6 — DB: 0.5 h. Evidence: 2.5 h (human elimination half-life, PMID 23099431). The DB value appears to be the distribution half-life (7.6 min) or a community estimate, not the elimination half-life.
2. BPC-157 — DB: 4 h. Evidence: <30 min (animal only — rats, dogs). No human PK data exists; the DB value contradicts all available animal evidence.
3. AOD-9604 — DB: 0.5 h. Evidence: ~3-4 min (animal IV). No human SC data; DB value is unsupported.

❌ Duplicate Record

4. Bremelanotide (PT-141) vs. PT-141 — Two records for the same compound with divergent frequency_hours (72 h vs 48 h) and slightly different descriptions. Must merge into one authoritative record.

⚠️ Ambiguous / Conflated Values

5. MK-677 — DB: 24 h. This appears to be the IGF-1 elevation duration, not the compound elimination half-life (4-6 h in beagles). The value conflates pharmacodynamic effect duration with pharmacokinetic half-life.
6. Semax — DB: 2 h. This is the downstream BDNF/neurotrophin effect duration, not the plasma half-life (~3-5 min).
7. GHK-Cu — DB: 0.5 h. Conflicting sources (30-60 min vs. 2-4 h); no definitive study.
8. Epitalon — DB: 4 h. Conflicting estimates (3 h vs. 6-8 h); no formal PK data.

⚠️ Minor Discrepancies (close but not exact)

9. Bremelanotide — DB: 2.5 h vs. FDA label 2.7 h
10. Tesofensine — DB: 216 h vs. 220 h
11. CJC-1295 with DAC — DB: 192 h vs. range 139-194 h (192 h is top of range)
12. Hexarelin — DB: 1 h vs. IV 75-120 min
13. SS-31 — DB: 1 h vs. animal ~2 h
14. 5-Amino-1MQ — DB: 7 h vs. community 4-6 h
15. Follistatin-344 — DB: 24 h vs. one source ~35 h

⚠️ Evidence-Limited (no human PK data, community estimates)

KPV, LL-37, DSIP, Dihexa, Pinealon, Thymalin, Selank, MOTS-c, NAD+, HGH Fragment 176-191, CJC-1295 w/o DAC, TB-500, PEG-MGF, IGF-1 LR3, Melanotan II — all lack human pharmacokinetic studies. Stored values are community estimates presented as established facts without caveat.

---

Proposed Corrections & Alignment Plan

Phase 1 — Schema Enhancement (no data changes)

Add metadata fields to Peptide.jsonc so half-life values carry evidence provenance:

half_life_hours          (existing — keep)
half_life_type           (enum: "elimination", "terminal", "distribution", "effect_duration", "estimated", "unknown")
half_life_unit           (string: "hours" — for future unit flexibility)
half_life_source         (string: citation/DOI/label reference)
half_life_species        (enum: "human", "animal", "in_vitro", "mixed", "unknown")
half_life_route          (enum: "sc", "iv", "im", "oral", "intranasal", "topical", "unknown")
half_life_confidence     (enum: "high", "moderate", "low", "evidence_limited", "unknown")
half_life_range_min      (number, optional — for compounds with a reported range)
half_life_range_max      (number, optional)

Phase 2 — Value Corrections (after approval)

Phase 3 — Duplicate Resolution

Merge "PT-141" (id 68e3282c...) into "Bremelanotide (PT-141)" (id 68c9c382...2f). Reconcile frequency_hours (72 h vs 48 h) per the FDA label (doses ≥24 h apart → 24 h minimum; 48-72 h community). Redirect any Injection/PeptideCycle/PlannedInjection references.

Phase 4 — Consumer Alignment

1. Consolidate decay model — All four inline implementations (PeptideDashboard.jsx, CurrentLevels.jsx, PeptideLevelsChart.jsx, peptideLevels.js) should import the shared estimatePeptideLevels() from src/lib/peptideLevels.js. The shared function becomes the single authoritative decay model.

2. Display metadata — All display surfaces should show confidence/evidence status alongside the value, e.g., "Half-life: 2.5 h (human, elimination, moderate confidence)" or "Half-life: unknown (evidence-limited)".

3. AI prompt context — ProtocolBuilder.jsx and CycleManagementModal.jsx should pass evidence metadata to the AI prompt so the model knows which values are established vs. community estimates.

4. Coach context — When peptide levels flow into Coach context, include confidence metadata so the Coach does not present evidence-limited estimates as established facts.

---

Evidence Sources

• FDA Prescribing Information (semaglutide, tirzepatide, tesamorelin, bremelanotide)
• Peer-reviewed PubMed (GHRP-6 PMID 23099431; ipamorelin PMID 10496658; CJC-1295 PMID 16352683; kisspeptin-10 PMC 3232613; IGF-1 PMID 8219484; DSIP PMID 6548966)
• Frontiers/PMC reviews (BPC-157 PMC 12313605; BPC-157 PK Frontiers Pharmacol 2022; SS-31 PMC 9192202)
• Clinical pharmacology (AOD-9604 JoFEM 2024; MK-677 Wikipedia/Chapman et al.)
• Evidence databases (Alzheimer's Drug Discovery Foundation for MOTS-c; PeptIQ, Peptidepedia, PeptideSchedule for research peptide community estimates)

All URLs verified as of 2026-10-22. Research peptide community sources are clearly distinguished from peer-reviewed evidence.

---

This audit is read-only. No peptide records, protocols, doses, logs, or calculations have been modified. Implementation awaits approval of the proposed corrections.