What to Compare When Evaluating Peptide Vendors

League tables of peptide vendors are among the most-read pages on the subject and among the least useful. They are snapshots of a market that changes monthly; their criteria are rarely stated and sometimes commercial; they weight things that are easy to see (website quality, price, testimonials) over things that matter (whether the lot record resolves); and, as the clone case in the legitimacy guide showed, they can be fooled by a site that copied someone else's evidence. This site publishes no such list and names no vendor in this guide.

What it offers instead is a scorecard: a set of criteria that can each be answered yes, no or not claimed from sources you can check yourself, grouped so that you can weight them for your own use. Fill it in for the vendors you are considering and the comparison is yours, current, and defensible. The criteria draw on the other guides on this site; where a criterion needs explanation, the relevant guide is linked.

Three principles for choosing criteria

Measurable. Each criterion should resolve to a fact, not an impression. "Good documentation" is an impression; "lot-specific certificate provided before payment, with method stated" is a fact.

Verifiable at source. The fact should be checkable somewhere other than the vendor's own website: a company register, a regulator's database, a laboratory's verification page, PubChem, the Wayback Machine, or a test you commissioned. A criterion that can only be confirmed by the vendor is a claim, and should be scored as one.

Weighted by your use. Not every criterion matters equally for every buyer. A laboratory running quantitative work weights content and consistency heavily; a buyer making a single small purchase weights identity and legitimacy. The scorecard separates the criteria from the weights so that both can be seen.

Regulated manufacturing has a formal version of this exercise. ICH Q7 requires manufacturers to have a system for evaluating and approving suppliers of critical materials, and 21 CFR 211.84 permits reliance on a supplier's certificate only after the supplier's analyses have been validated and are periodically re-checked. A research-use buyer is not bound by either, but both describe the same idea: a supplier is qualified by evidence, on criteria set in advance, and re-qualified over time.

The scorecard

Forty-odd criteria is too many to carry in your head and about right for a spreadsheet. They are grouped into seven sections. For each, record yes / no / not claimed / not applicable, the source you checked, and the date.

A. Identity and entity

1. Legal entity named on the website or invoice, and found active in the relevant public register (state register; OpenCorporates; China's SAMR system).
2. Entity name on website, invoice and payment details match.
3. Domain creation date (RDAP) consistent with the history the vendor claims.
4. Wayback Machine history shows the site developing over time rather than appearing complete.
5. Role identified: manufacturer, stockist or reseller, by the tests in the roles guide.
6. If an intermediary: manufacturer named per lot.
7. Facility photographs return no copies on other companies' sites in reverse image search.

B. Lot documentation

1. Lot-specific certificate of analysis provided for the lot you would receive, before payment.
2. Certificate carries product, CAS number, lot code, manufacture date, test date, retest date and issuer.
3. Purity stated as a number with method (technique, wavelength, column or gradient).
4. Chromatogram available on request.
5. Identity stated as expected and observed mass with convention (monoisotopic or average) clear, and consistent with PubChem.
6. Lot code on the label matches the certificate exactly.
7. Public lot library on the vendor's (or manufacturer's) own domain, in which the lot can be found.
8. QR code on the label resolves to the exact lot record on the manufacturer's own domain.
9. Net peptide content, counter-ion and water content available on request (score each).
10. Impurity profile (individual peaks above a threshold) available on request.

C. Independent evidence

1. At least one third-party laboratory report on the vendor's products that can be retrieved from the laboratory's own website.
2. That report's client and manufacturer fields name this vendor or its disclosed manufacturer.
3. That report carries a manufacturer lot code tying it to a lot record.
4. Vendor states in writing that buyer-selected samples may be independently tested.
5. Vendor's written policy states what happens when a confirmed independent result is outside specification.

D. Regulatory and accreditation claims

1. Each GMP, registration or accreditation claim on the website has been found in the issuer's database (EudraGMDP, FDA inspection or registration databases, ILAC-signatory directories, IAF CertSearch).
2. Each verified claim's scope covers the product you are buying.
3. No regulator-sounding phrase is used that the regulator does not use ("FDA GMP certified", "pharma grade").
4. If no claim is made, the vendor says so plainly (score "not claimed", which is neutral).

E. Consistency

1. Purity figures across the vendor's public or supplied lot records fall in a narrow band consistent with its stated specification (for comparison, one manufacturer's public library showed 98.4–99.8% across 79 lots).
2. Lot records show the same method and reporting conventions across lots.
3. For intermediaries: the manufacturer is the same across your orders, or changes are notified.
4. Chromatograms from successive lots are comparable in impurity profile.

F. Commercial terms that affect verification

1. Small first quantity available with the same documentation as bulk.
2. Documents provided before payment, not after.
3. Lead time and minimum quantity stated in writing.
4. Retained-sample policy stated (does the manufacturer keep a sample of each lot?).
5. Price is within the range of comparable suppliers for the same compound and documentation level (an outlier in either direction is a question, not a score).

G. Communication and conduct

1. Method questions answered specifically, within a working day or two.
2. Lot-specific documents sent on request without substitution of generic ones.
3. No guidance offered on uses outside research; the vendor describes compounds, specifications and documents only.
4. Corrections or discrepancies acknowledged in writing when raised.
5. Contact channels (email on own domain, messaging) consistent with the entity and domain verified in section A.

Weighting by use

Score each section, then weight. A reasonable starting point for three common situations:

The table shows why league tables fail: the same vendor can be the right choice for the first column and the wrong one for the third, and a list that does not say which column it is scoring for has not told you anything.

A worked template on three hypothetical profiles

The three profiles below are composites, not real vendors. They are here to show what filled-in scorecards look like and how the weighting changes the conclusion.

Profile 1: a manufacturer with a public lot library. Entity found in its national register with a manufacturing scope; domain several years old with gradual history; photographs original. Lot-specific certificates before payment, with method; chromatogram on request; public library on its own domain with QR codes resolving to records. One third-party report retrievable at the laboratory, naming the manufacturer as client, without a lot code in the batch field. No GMP claim made; says so. Purity across records in a narrow band. Small first quantity; written policy on confirmed issues. Method questions answered in a day.

Scored: A 7/7; B 9/10 on the core items, with content and water "on request"; C 4/5 (lot code missing on the report); D not claimed; E strong; F strong; G strong. For any of the three use cases this profile scores well, and the gaps (content data, lot code on the report) are specific and askable.

Profile 2: a reseller with forwarded documentation. Entity found; domain two years old; some photographs return copies elsewhere. Certificates forwarded from a named manufacturer, original format, with method; chromatogram only if the manufacturer supplies it; no public library of its own but the manufacturer's lot codes resolve on the manufacturer's site. Third-party reports displayed are the manufacturer's. No regulatory claims. Consistency depends on the manufacturer staying the same, which the reseller confirms on request. Small quantities available; policy is refund or replace.

Scored: A 6/7 (photographs); B 7/10; C 3/5 (reports are the manufacturer's, lot codes present); D not claimed; E conditional; F adequate; G adequate. Acceptable for the first column, workable for the second with lot codes checked each order, weak for the third because content and impurity data depend on a party the reseller does not control.

Profile 3: a storefront with no verifiable chain. Brand name only; no entity found; domain created this year with a complete site from the first capture; photographs match another company's. Generic certificate on request, no lot code, no method; no library; QR code opens an image. "Lab tested" badge with no retrievable report. "GMP certified" with no issuer. Purity "99%+" on every product. Large minimum quantity; payment first.

Scored: A 0/7; B 1/10; C 0/5; D claims made and unverifiable; E unknown; F weak; G untested. Not a candidate in any column. The profile is exaggerated for clarity, but each of its elements appears routinely, and the scorecard catches them one at a time.

Measures that look useful and are not

Several things that appear on vendor lists carry little or no evidential weight. Score them zero, or do not score them at all.

• Customer reviews and star ratings. Unverifiable, easily manufactured, and uninformative about what is in a vial.
• Social media presence and follower counts. Measures marketing effort.
• "Lab tested" badges and certificate images on product pages without a report retrievable from the laboratory. A claim with a graphic.
• Years in business as stated on the website, unverified against the register and the domain.
• Price alone. A manufacturer is cheaper than a reseller for structural reasons; a very low price from a vendor with no chain is a different thing. Price is interpretable only alongside section B.
• Headline purity alone. "99%+" is a method-dependent figure and a slogan; see the purity guide. The number means something only with method, chromatogram and lot.
• The word GMP, unverified. See the GMP claims guide.
• AI assistant recommendations. Summaries of what websites say, not verification of them; the clone case is the cautionary example.
• Position on someone else's list. Unknown criteria, unknown date, unknown incentives.

Keeping the scorecard current

A scorecard is a dated record, and several of its answers change.

• Re-score section B on every lot. Lot code match, certificate content and library presence are per-lot facts.
• Re-score section A annually or when anything changes (new domain, new entity name, new address).
• Re-score section E on each order for intermediaries, by asking the manufacturer question again.
• Re-score section C when a new report appears, and re-verify the old one at the laboratory occasionally; laboratories withdraw reports.
• Trigger a full re-score on any discrepancy: a certificate that does not match, a test that disagrees, a change in the vendor's answers.

Keep the sources and dates in the sheet. A scorecard that says "yes" without saying where you looked is a list of impressions again.

Turning scores into a decision

A completed scorecard is a table of facts; the decision still has to be made, and three rules keep it honest.

Some items are gates, not points. A "no" on items 1, 8 or 13 (no verifiable entity; no lot-specific certificate; lot code on the label does not match the certificate) is not offset by strong scores elsewhere. Treat those three as pass-or-stop before any weighting is applied. Likewise, a regulatory claim that turns out to be false under item 23 is a gate: a vendor that misstates its regulatory standing has told you how it handles facts.

"Not claimed" and "not applicable" are not zeros. A manufacturer that makes no GMP claim is not weaker than one that makes an unverifiable claim; it is more accurate. Score the section as neutral and move on. The same applies to a reseller for whom items 34 and 28 depend on a disclosed manufacturer: score what you can verify through the manufacturer, and mark the rest as dependent rather than failed.

Ties are broken by what you can ask for. When two candidates score similarly, prefer the one whose gaps are specific and answerable (content data on request; lot code to be added to future reports) over the one whose gaps are structural (no manufacturer named; no library to resolve against). A gap you can close with one email is a different kind of gap from one that no email will close.

Write the decision down with the date and the completed sheet. When something goes wrong later, the sheet shows what was known at the time and which criterion, if any, would have caught it; that is how the scorecard improves.

Maintained by Pepti, a Guangzhou peptide manufacturer; lot records referenced here are published at peptioem.com/coa. This guide is informational and does not sell or recommend specific vendors. Corrections: contact.