How Peptiter handles evidence, machine learning, and prescriber autonomy.
We are a clinician decision-support platform — not a drug company, not a pharmacy, not an autoprescribing AI. These eight principles bind every product release, every model update, and every commercial agreement.
Evidence tiers are explicit, never hidden
Every recommendation surfaces its evidence tier (A: RCT, B: cohort/registry, C: mechanistic/case series, D: in-silico). Prescribers see the tier before they see the protocol.
ML is gated, not generative
Models do not invent peptides or doses. They re-rank within a curated catalogue and trigger only when ≥50 prescriber decisions, ≥3 evidence-tier sources, and statistical confidence thresholds are met.
Prescriber is the final node
Peptiter is FDA CDS Exempt under §520(o)(1)(E) precisely because the licensed prescriber independently reviews every recommendation. No autoprescribe path exists.
AKS-clean economics
Flat monthly licence. Zero per-prescription fees. Zero pharmacy referral fees. The platform earns nothing from any specific drug or pharmacy choice.
Continuous data, continuous audit
CGM, BHB, RR, and questionnaire ingestion is timestamped, hashed, and append-only. Every score (OII, CIR) is reproducible from raw inputs.
DiscoverRx publishes the gates
The 10-gate computational pipeline (Boltz-2, PRODIGY, PROSPERousPlus, ToxinPred3, ChEMBL calibration) is public. Pass/fail thresholds are versioned. No black-box scoring.
Patients can request triangulation
Subjective questionnaires (PHQ-9, GAD-7, ASRS, PSQI, AUDIT-C) are weighted alongside objective signals and labs. PHQ-9 Q9 triggers automatic 988 escalation.
We invite scrutiny
Peptiter operates under IRB-determined protocols and welcomes external researchers. Pre-registered comparisons with current standard-of-care are the goal, not press releases.