Written for prescribing physicians, regulators, distributors, and serious buyers who need to scrutinise our process before relying on a rating. What we do, how we do it, what our scores can and cannot tell you.
Version 1.0 · August 2026 · Changes to this document are versioned, dated, and announced to registered clinicians and members.
200+
Suppliers in dataset
32
Publicly rated
100%
Ratings human-signed
0
Commissions taken
Scope of assurance
PepiScreen evaluates the evidence a peptide supplier publishes — laboratory reports, chain-of-custody documentation, batch traceability, and public transparency signals. Ratings assess demonstrable practice; they are not a certification of product safety, efficacy, or medical suitability. Every rating answers a single question for the reader: how much of what this supplier claims can be independently substantiated from their own paperwork?
The rubric — fixed weights, published rules
Every supplier is scored against the same four-part rubric. Weights have not moved since publication; any future adjustment carries a versioned changelog with recomputed historical scores retained.
Independent lab
Third-party, ISO 17025-accredited laboratory named on every certificate, with verifiable contact details and test IDs.
30%
Analytical evidence
HPLC chromatograms, LC-MS or MALDI-MS confirmation of identity, residual solvents, water content, and peptide-identity checks.
25%
Batch traceability
Every certificate ties back to a real batch number, manufacturing date, test date, and retest/expiry date.
20%
Additional transparency
Business registration, GMP evidence, ISO accreditation, storage & shipping standards, customer redress, historical consistency.
25%
Total score is on a 0–100 scale. Band cut-offs (HIGH ≥ 70, MEDIUM 45–69, LOW 25–44, AVOID < 25) are the same for every supplier.
How AI accelerates review — and what humans still decide
Our proprietary AI-assisted pipeline lets a small independent team do the ingestion and cross-reference work that would otherwise take dozens of individual private testers. It does not replace the judgement that produces the final rating.
Where AI does the work AI
OCR and structured field extraction from COAs and lab reports (approximately 15 fields per document)
Cross-referencing named laboratories against a maintained ISO 17025 accreditation registry
Duplicate detection across the whole dataset — matching batch numbers, test dates, and chromatogram fingerprints to surface recycled or copied certificates across brands
Molecule-identity checks against reference peptide sequences and known impurity patterns
Final rubric score sign-off on every published rating
Weight adjustments in edge cases, with the rationale published in the audit note
Escalation review on every red flag AI surfaces
Response to supplier rebuttals, re-review requests, and community complaints
Charter enforcement decisions on member suppliers
Every published rating carries a reviewer's initials and the date the review was signed off. Ratings that AI processes flagged but a human reviewer overrode carry a visible note explaining why.
The evaluation flow
Ingest. Supplier COAs, product pages, batch records, and lab attestations are pulled or uploaded. Documents are hashed on receipt so we can prove later exactly what we scored.
Extract. AI parses each document into structured fields (supplier, batch #, test date, laboratory, purity %, water %, residuals, endotoxin, sterility, molecule identity, expiry).
Cross-reference. Every field is checked against internal registries (ISO 17025 accredited labs), across our full dataset (duplicate COAs, shared batch numbers), and against reference peptide chemistry.
Score. The deterministic 30/25/20/25 rubric produces a raw 0–100 score; band assignment is automatic.
Human review. A PepiScreen reviewer verifies flagged items, sanity-checks the score against evidence quality, adds the audit note, and signs off. Nothing is published without this step.
Publish. The score, strengths, weaknesses, missing evidence, and reviewer sign-off appear on the supplier's directory listing with a link back to the source documents that produced each rubric point.
Continuous audit. Quarterly re-review, community complaint intake, and dataset-wide cross-checks may downgrade or upgrade a rating at any time. History is retained.
What our scale changes
Individual private tests — one buyer sending one vial to one lab — answer one question about one batch. A rating drawn from cross-referencing against a register of 750+ suppliers answers a different, harder question: does this supplier's pattern of evidence look like the patterns we see from the ones already verified?
Concretely, the scale lets us detect what no single-supplier private test can:
Recycled COAs surfacing across two, three, or four “different” brands — catches white-labelling that private testing never sees.
Laboratory-shopping: suppliers who use different labs each quarter to avoid unfavourable retests.
Clusters of suppliers demonstrably sharing the same upstream Chinese manufacturing facility, disclosed or not.
Batch-number recycling within a single supplier across time.
Real-world purity-versus-price benchmarks per peptide, updated weekly, that make outlier claims fall out of the distribution.
We currently maintain 32 publicly rated suppliers (20 full-audit + 12 public-evidence screened), 10 country-specific prescription-law summaries, and 100+ additional providers logged and awaiting evaluation. Each figure grows every week.
Data mapping
Every field a rating depends on is stored with a source hash and a click-through to the exact document the rubric point cites. When a physician opens a supplier's listing, no claim on that page rests on our word alone — each of the six rubric outcomes links to the source COA page, lab attestation, or batch record that produced it.
Reliability and explicit limits
What our score CAN tell you
How much of what a supplier claims is substantiated by documentation an independent third party can verify.
Whether the supplier's certificates have been forged, edited, or recycled from another brand's paperwork.
Whether the supplier operates a batch-traceability regime a physician or importer could audit.
How the supplier's transparency and evidence compare to a register of 750+ peers (with detailed evidence data on the growing subset who have submitted certificates).
What our score CANNOT tell you
Whether a specific batch is safe to use or clinically appropriate for a given patient.
Whether the supplier will remain compliant with the Charter in future.
Whether a product is legally importable in your jurisdiction (see /doctors/ for country-law notes).
The therapeutic effect, dosing, or medical suitability of any peptide — those decisions belong to the physician.
Independence & funding — structural, not marketing
PepiScreen does not sell peptides. It does not sell laboratory testing. It does not take commissions on any supplier's sales, and it does not vary scores based on paid membership. Charter membership and Influencer subscriptions fund the review infrastructure; scores are produced independently of any member's status. Reviewers do not know which suppliers are paying members at the time the rubric is applied. Nothing on this site is an advertisement placed by a supplier.
This funding structure is the reason a commercial lab-testing brand cannot honestly replicate what we do. A lab that sells tests to suppliers has revenue riding on those same suppliers' good will — the identical conflict a supplier has when it self-tests. Our funding does not rest on any supplier's cheque. Removing that conflict is the point.
Change log & version control
Version
Effective
Change
1.0
August 2026
Initial publication. Four-band scale (HIGH / MEDIUM / LOW / AVOID). Two audit tiers (Full audit, Screened).
All material changes are versioned, dated, announced to registered clinicians and members, and reflected in a public changelog. Historical scores are retained; recomputed scores under any new methodology carry a version tag.
Contact for scrutiny
Physicians, regulators, or researchers wishing to audit our process, request a walk-through of a rating, or challenge a specific supplier's score: our contact form. We publish every substantiated correction, credited or anonymous as the correspondent prefers.