How We Rate Peptide Vendors

The Short Version

Every vendor gets a score out of 10 and a tier label. The score combines five things we look at, weighted by a fixed formula. One rule, the testing gate, can hold a score down when fewer than half of the vendor's products have a certificate we confirmed with an independent laboratory.

Every rule below is applied to every vendor in the same way, with no exceptions, and every point value is printed on this page with the reason for it. The code that computes the scores is checked against this page automatically before every update of the site, so the page cannot drift from the scoring. This is version 2 of the rules, in force since 22 October 2026 (see the changelog at the end).

We rate vendors against each other within the research-peptide market, not against pharmacy-grade pharmaceutical companies. An Excellent vendor is best-in-class within this market; a Fair vendor is functional, with gaps that are common in this market.

The Five Things We Look At

Each vendor is scored 0 to 10 on each dimension, then the five sub-scores are combined using these weights:

The overall score is the weighted sum: lab × 0.55 + shipping × 0.20 + pricing × 0.15 + operating history × 0.05 + transparency × 0.05. A vendor is scored only once all five sub-scores are in place.

Rounding. Each sub-score is rounded to one decimal place (half up). The overall score is computed from the rounded sub-scores, then shown to two decimal places, rounded half up on the exact value. Tiers, ranks and ties all use the score as shown.

What the Tiers Mean

Scores are anchored to the US research-peptide market as it actually is, the set of vendors a buyer can realistically choose from, not to a pharmaceutical-grade ideal that no vendor in this market meets.

1. Excellent8.50 to 10 / 10

   Best-in-class within this market. Certificates from independent laboratories for nearly everything on sale, many of them confirmed with the issuing laboratory, alongside clear policies and a solid operating record.

2. Strong7.00 to 8.49 / 10

   Independent testing across most of the range and clear policies, missing one or two of the strongest signals (for example, endotoxin and sterility testing, or certificates we could confirm with the laboratory).

3. Fair5.50 to 6.99 / 10

   Functional, with the gaps that are common in this market: partial certificate coverage, thinner documentation, or a shorter operating record.

4. Weak4.00 to 5.49 / 10

   Limited testing evidence for the products on sale, combined with weaknesses elsewhere such as restrictive refund terms or little information about the business.

5. Poorbelow 4.00 / 10

   Little or no third-party testing we could verify, or several serious gaps.

A tier describes a range of scores, not a quota: if no vendor currently reaches a tier, that tier is empty.

How the Lab Sub-Score Is Calculated

The lab sub-score is built product by product: each product is scored from the certificate of analysis on file for it, the scores are averaged, and the adjustments below are taken off.

A1. Which products count

• The products in our catalogue that the vendor sells, excluding lab supplies such as bacteriostatic water. This is our catalogue of peptides, not every product the vendor sells.
• A product with no published certificate counts as 0, so a vendor that publishes no certificates at all has a lab sub-score of 0.0.
• A certificate for a different size of the same product counts for that product. A certificate for a blend counts only for that blend, not for the single peptides in it.

A1b. Core products and niche products

• The core products are the 15 products in our catalogue carried by the most vendors (ties broken alphabetically). The list is computed from our listings, not chosen by hand, and each vendor page marks which of its products are core.
• A missing certificate counts fully against a vendor for a core product, and at half weight for any other product. Every product that has a certificate, including a flat 2.0 or a zero-scored one, counts fully.
• Why: the core products are what most buyers order, so a gap there matters most. A vendor that offers a wider range should not lose as much for an untested niche product as for an untested staple, but an untested niche product is still a real gap for the buyer who orders it, so it still counts. Without this, a small catalogue could outscore a broad one simply by being small. Half weight sits honestly between counting fully and not counting; it is a judgement, and it is stated here so anyone can redo the arithmetic.

A2. Each certificate is scored like this

• A published certificate starts at 3. Why: a certificate that names a laboratory is evidence someone tested something; the points below say how much it can be trusted.
• +2 if the issuing laboratory is in our lab directory at tier A+, A, A-, B+, B or B-. A tier C laboratory earns 0 here. Why: tiers B- and above are laboratories we could establish as real and independent with some public record; a tier C laboratory offers no way to check a report and shows neither identity by mass spectrometry nor a track record, so there is too little to credit.
• +1.5 if the certificate carries a verification key (the report number or access code the laboratory issues for that test). Why: a key lets anyone, not only us, check the report with the laboratory.
• +1.5 if we confirmed that key at the laboratory's own site, through its verification portal or its hosted copy of the report. Why: confirmation is the strongest evidence a published certificate can carry, that the laboratory itself stands behind this report; it is worth as much as printing the key because it is the check the key exists for.
• +0.5 if the certificate includes an endotoxin result. +0.5 if it includes a sterility result: USP <71>, or a PCR or rapid sterility screen. Microbial limits (total aerobic and yeast and mould counts, TAMC and TYMC, or USP <61> and <62>) are not a sterility test and earn nothing. Why: these tests matter for anything injected, but they are worth less than identity and purity because a certificate without them still tells you what is in the vial.
• +1 if the laboratory is accredited, as shown on the accrediting body's public register, with a scope that covers peptide testing (check L6 under How We Rate the Labs Themselves; a laboratory's own claim or a logo on a certificate does not count). Why: accreditation means outside assessors checked the laboratory's methods and equipment.
• Then the purity adjustment below, with the result kept between 0 and 10.

Purity adjustment, where the certificate reports an exact purity figure:

• 99.7% or higher: +0.5.
• 99.0% up to but not including 99.7%: no adjustment (the usual level in this market).
• 97.0% up to but not including 99.0%: −1.5.
• Below 97.0%: −3.0.

A certificate that states only a specification (such as "≥99%") rather than a measured figure gets no purity adjustment either way. Why these bands: 99% is the usual research-grade figure; a result well below it means a meaningful share of the vial is something else.

A2b. Single-test reports. A report that tests only endotoxin, only sterility or only heavy metals is not a certificate of identity or purity and is never scored on its own as a product's certificate. An endotoxin-only or sterility-only report adds its +0.5 to the product's scored certificate when it covers the same product (any batch) within the last 12 months.

A3. Flat 2.0 when the source cannot be established

A certificate scores a flat 2.0, with no points added or taken off, when it is self-issued by the vendor, when it names no laboratory, or when the laboratory's independence is not established (tier D in our directory). Why: such a document may be accurate, but nothing outside the vendor stands behind it, so it is worth less than any certificate from a laboratory we could establish, and a little more than nothing.

A4. Certificates that score 0

A product's certificate scores 0 when:

• there is no certificate for the product;
• the laboratory's own record reports a FAIL (identity, purity or specification);
• the certificate itself reports a failed result (identity, purity or specification), including when its key is not found at the laboratory; or
• the copy the vendor posts differs materially from the laboratory's own copy: a different product, quantity, purity, client or test date. Formatting, re-saving a file, or a hyphen or typo in a lot label is not a material difference.

Why: a certificate the laboratory failed, one that reports its own failure, or one that does not match the laboratory's record, is no evidence for the product as sold.

A5. A key printed but not found at the laboratory

When a key printed on a certificate is not found at the laboratory (we try every letter case and the laboratory's own lookup, and retry), the certificate keeps its +1.5 for the printed key and earns no +1.5 for confirmation. There is no other deduction. A server error, a timeout or a bot challenge is never treated as "not found", and a laboratory that lists only the reports a client chooses to publish is treated as "not confirmed", not "not found". Keys on certificates that are not the scored certificate for a product we list are not scored at all.

A6. Age

A certificate's date is the date the laboratory received or tested the sample, as printed. If no date is printed, it is the earliest date we can prove the certificate existed (file metadata, an archived capture, or our first recorded audit of it). A certificate with no date at all is treated as older than 12 months.

When a vendor posts more than one certificate for a product, across sizes or lots, the product is scored from the newest certificate by this date. Why: the newest certificate is the one most likely to describe current stock, and choosing it removes any choice of the best-scoring copy.

For each certificate more than 12 months old, the lab sub-score loses 0.25 × that certificate's score ÷ the weighted number of products (A1b). Why: an old certificate may not describe current stock, so a quarter of its weight comes off; a vendor with recent certificates loses nothing.

A7. Adjustments

Only three adjustments exist, each with a fixed amount and a cap:

CLAIM is never used for products that merely lack a certificate: coverage already scores those 0. Why these amounts: each describes a problem with how the evidence is presented rather than with a single product's result, so each is small next to a certificate's own score, and capped so it cannot outweigh the certificates.

Shared analyses. When the same laboratory analysis, with the same dates, results and chromatogram, appears on reports issued to two vendors, the certificate still counts and SUPPLIER does not apply as long as the report names the vendor as client. We state the match on each vendor's review, so readers know the two sets of results are not independent tests.

A8. The result

The lab sub-score is the weighted mean of the product scores (weights from A1b), minus the adjustments, floored at 0 and rounded to 0.1. Each vendor page shows this working product by product.

How We Check Certificates

Most peptide vendors say "third-party tested." We do not take that on faith. For every rated vendor, we:

1. Find the vendor's certificate index, lab-reports page, or per-product certificate links.
2. Download and read the certificate for every product we list.
3. Match the named laboratory against our lab directory.
4. Record the verification key, the report number or access code the laboratory issues for each test, and, where the laboratory offers a lookup, confirm it there.
5. Record the tests actually run (identity, purity, endotoxin, sterility, microbial), plus the batch number and test date.

On product and vendor pages, the certificate column links to the certificate the vendor publishes. Where we have read and checked it, or confirmed it with the issuing laboratory, the link is marked ✓ Verified.

Honest limits. Some vendor sites block automated access, and some are offline when we check; those are re-checked by hand or at the next audit. Confirming a certificate tells you who tested the sample and what they found. It does not tell you that the vial you receive came from the same batch.

Evidence Tiers

"We verified the certificate" can mean very different things. To make the difference visible, every carried product is placed in one of five evidence tiers, rolled up per vendor on the verification report.

Key found at the lab

We found the certificate's key at the laboratory that issued it, through the lab's own verification portal or its hosted copy of the report. This tier records the check we made, not the certificate's score: a certificate whose key was found can still score a flat 2.0 (A3, for example when the laboratory is tier D) or 0 (A4, when the laboratory's record reports a FAIL or the vendor's copy differs from it). Only a certificate the lab rule scores on the full rubric (A2) is called lab-confirmed or marked ✓ Verified anywhere on the site.

Read & checked

We read the certificate ourselves, or confirmed it on a platform that signs results on a laboratory's behalf, and checked its lab, batch and purity, but could not confirm it with the issuing laboratory itself.

COA linked

The vendor links a certificate that we have not yet read or confirmed.

Lab claimed

A laboratory is named, but no certificate for the product is published.

No COA

Neither a certificate nor a laboratory is published for the product.

The verification page turns these tiers into a 0 to 100 evidence score per vendor (key found at the lab = 4, read and checked = 3, linked = 2, claimed = 1, none = 0, divided by the maximum possible). The evidence score is a summary for readers; the lab sub-score above is what feeds the vendor's rating.

How We Rate the Labs Themselves

Every laboratory in our lab directory is graded by the same six checks, only from evidence we can cite with a link and the date we checked it. What we could not check is recorded as not checked, never assumed. Each laboratory's profile shows the result of every check, its evidence link and the date.

The six checks

• L1 Real business. An entry in a state or national business registry for the operating company, or a physical street address shown by the laboratory (not a mailbox or virtual office) together with named staff and their roles. Why: a certificate is only worth something if an identifiable business that a reader or a regulator could reach stands behind it.
• L2 Independence. All four must hold: (a) the laboratory's domain was registered before the earliest report we hold from it; (b) its reports name more than one unrelated client, or it publicly offers testing to the public; (c) it has no ownership, address, staff or hosting link to a vendor we list; and (d) its lookup does not run on a vendor's site or code. If any of them fails, independence is not established. Why: a laboratory that is in effect one vendor's in-house operation cannot vouch for that vendor's products as a third party.
• L3 Verification route. A public lookup, or a copy of the report hosted by the laboratory, that a reader can use for a specific report and that returns the report for a real key and rejects an invented one. We test both on the day we check. A library that lists only the reports a client chooses to publish, or reports sent only on request, counts as partial. Why: this is what lets anyone, not only us, check a certificate with the laboratory instead of trusting the vendor's copy.
• L4 Identity method. Identity by mass spectrometry (LC-MS or MS) passes. Identity by HPLC retention time alone, or no identity test, is partial. Why: mass spectrometry measures the molecule itself; a retention time only shows that something came off the column where the reference did, which another compound can also do.
• L5 Track record. At least 12 months between the earliest dated report we can find from the laboratory (for any client) and the day we check, and at least three unrelated clients. Why: a laboratory with a history across several clients has more to lose from a wrong report than one that appeared recently for one or two.
• L6 Accreditation. ISO/IEC 17025 accreditation shown as active on the day we check on the accrediting body's own public register (A2LA, PJLA, ANAB, IAS, UKAS, DAkkS or similar), with a scope that covers peptide testing: HPLC purity or mass-spectrometry identity of peptides, or endotoxin for a laboratory that tests only endotoxin. A laboratory's own claim or a logo on a certificate does not count, and suspended, withdrawn or expired accreditation is not accreditation. Why: accreditation means outside assessors checked the laboratory's methods, but only the methods in its scope; accreditation for other work says nothing about a peptide result.

The tier follows from the checks

The first row that applies is the laboratory's tier.

How a tier reaches a vendor's score. Only through the certificate rubric above: rule A2 gives +2 to a certificate from a laboratory at tier A+, A, A-, B+, B or B- (B- or better), nothing for tier C, and rule A3 scores every certificate from a tier D laboratory a flat 2.0. Tiers B- to A+ earn the same +2 laboratory points; the tier letter shows how much of the laboratory's work a reader can check. The +1 accreditation point (A2) is given exactly when L6 passes, whatever the tier, and only to a certificate tested on or after the date from which the accrediting body's register shows that accreditation as active. A laboratory's own claim does not count. A tier D laboratory's certificates still score the flat 2.0 with nothing added. Why: the checks ask what a reader most needs to trust a certificate (a real business, independence, a way to check the report) before what makes it better (identity by mass spectrometry, a track record, accreditation), so a missing basic outranks any refinement.

Clarifications

• L0 Every laboratory is graded. A laboratory is graded as soon as it is first named anywhere in our records, on a certificate or in a vendor's claims, using its own public reports for any client. No laboratory named in our records is left without a tier.
• L1 Registry match. When a laboratory's own site names no company, L1 can still pass on a business-registry entry that is the only match for the laboratory's name in the state or area the laboratory gives, and whose registered purpose is laboratory or analytical testing. The match and why it is unique are recorded on the laboratory's profile; a generic name with several possible matches does not pass.
• L1 Linked registry entry. L1 also passes on a business-registry entry tied to the laboratory by a named staff member's public professional profile naming that company as employer, or by a job posting naming both the company and the laboratory. The link and its sources are recorded on the laboratory's profile.
• L2b Clients. A laboratory passes L2 (b) when its reports name more than one unrelated client, or when it publicly offers testing to the public (an accrediting-body register entry, a commercial services and pricing page, or a long registry history of testing work). A laboratory whose only evidence of clients is one vendor, with nothing showing it serves the public, fails (b).
• L2c Shared ownership. A link to a vendor we list (ownership, address, staff or hosting) does not by itself fail L2 (c) when the laboratory serves unrelated clients, passes L1 to L4, and passes a documented deep check: its people, registry records, analytical consistency of its reports, and agreement with other laboratories on the same lots. If the link is proven and any of those fail, L2 fails and the laboratory is graded D. A proven link is disclosed on the laboratory's profile and the vendor's page, with evidence and date.
• L2a/L5 Dates. For L2 (a) and L5, a single report date that the laboratory's own records contradict (its neighbouring report numbers, the file's creation data, or a corrected copy) is treated as a misprint, documented with the evidence, and not used. Every other date is used as printed.

Names on certificates. A certificate is matched to a laboratory in our directory only when the laboratory name it gives is exactly the laboratory's name or a spelling recorded on its profile (ignoring capital letters), never by a partial match, so a similar-sounding name never takes on another laboratory's tier.

See the lab directory →

The Testing Gate: A Cap When Testing Is Not Verified

A9. The gate is computed, not set by hand. A vendor counts as third-party tested when at least half of its scored products (A1) have a certificate from a laboratory in our directory that is not tier D and not self-issued, with the verification key confirmed at the laboratory. Otherwise the gate applies. A certificate scored 0 under A4 does not count. Each vendor page shows the count.

The gate records that testing could not be verified to that standard; it is not a finding of a quality or safety problem. Why half: it asks that confirmed testing covers most of what a vendor sells, without punishing a vendor for a few products whose certificates are not yet confirmable.

When the gate applies, the vendor's score is capped. The cap is not a flat number: it starts at 6.0 and rises with operating history, shipping and consumer reviews.

A gated vendor's displayed score is the lower of its weighted score and its cap. If the weighted score is already below the cap, the displayed score is unchanged and the gate is informational. Either way, the vendor page says that third-party testing was not independently verified. Why the cap rises: a long record, good recourse and many good reviews are some reassurance when testing cannot be confirmed, but never enough to reach the top tier.

About Consumer Reviews

Where we hold a customer-review sample for a vendor, it is not shown on the site; it is used only in the testing cap above, adjusted for sample size. Five glowing reviews do not count for more than a hundred mixed-but-positive ones.

The adjustment is a technique called Bayesian smoothing: the average is pulled toward 3.5 (the value we assume for an unknown vendor) as if ten extra 3.5-star reviews were added. Smoothed average = (average × count + 35) ÷ (count + 10). With few reviews the result stays close to 3.5; with many, the vendor's own average takes over.

Example. Three 5-star reviews smooth to about 3.85, not 5.0. 200 reviews averaging 4.5 smooth to about 4.45, barely moved, because there is enough data to rely on.

A Worked Example

Say "Acme Peptides" gets these sub-scores from us:

If fewer than half of Acme's products had a lab-confirmed certificate, the gate would apply and its cap would be: 6.0 base + 1.0 (operating history sub-score 6.0) + 0.5 (shipping 6 or higher) = 7.5 / 10. Since Acme's score is 7.40, the cap does not reduce it, but the vendor page would still say that third-party testing was not independently verified.

Rankings, Ties and Product Pages

Ties. Vendors with the same score to two decimal places are listed alphabetically by name. Why: a difference below one hundredth of a point is not meaningful, and alphabetical order favours no one we choose.

Product pages. Ordered by the lab-rule score of each vendor's certificate for this product (out of 10), then by the vendor's published score, then alphabetically. The certificate score is the same rule (A2 to A6) that builds each vendor's lab sub-score. Price per milligram is shown but does not change the order.

Stores Not Taking Orders, Pending Updates and Signed-In Reads

F1. A store not taking orders. When a store says it is not accepting orders, or its domain shows a parking page, the vendor page shows a "Not taking orders" badge with a dated notice, shop links are paused, and we check the store daily for a reopening. The vendor keeps every sub-score and its testing gate from its last full audit: it is not re-scored from missing pages, and the closure does not change its gate. It is re-audited when it reopens. A vendor is delisted only when it stops selling for good, stops selling directly, or no longer sells to US buyers.

F2. Updates pending. At our editorial discretion, and on the same terms for every vendor, we may show a dated notice when a vendor tells us it is making changes we will be able to check, such as new testing or revised policy pages. The notice records the vendor's own account, which we have not verified; it appears above the score and as a label in the rankings; it never changes a score or rank and does not pause the review; and it lapses 30 days after its date if no re-audit has happened. A policy the vendor says it is changing is scored as it stands today.

F3. Vendors read through a signed-in account. Some vendors show their catalogue, prices or certificates only to signed-in customers. We read those through an account the editor holds, and score them exactly like any other vendor; a product whose certificate we still could not obtain scores 0. The vendors read this way are listed here: Felix Chemical Supply, Limitless Biotech, GenX Bio, Onyx Biolabs, Research Peptides Lab, Peptide Warehouse, Modern Aminos, Peptide Labs, VoriLabs, OMNI Peptides, Luv Aminos, Southern Aminos, Peptide Pro.

Corrections and Removal Requests

If something on a vendor page is wrong, tell us through the Submit page or by email, with the evidence. We check it, and if we were wrong we correct the page and record the change in the vendor's audit history. A vendor that publishes new certificates, changes laboratories or fixes a policy can ask for a re-audit at any time.

Reviews are not removed on request, including after a disputed finding. A vendor is delisted only when it stops selling for good, stops selling directly, or no longer sells to US buyers.

What We Don't Measure (and You Should Check Yourself)

• Whether peptides are right for you. That is between you and a qualified medical professional.
• Legal status in your state or country. Rules vary and change. Verify before ordering.
• Specific batch results. We rate the vendor's published evidence. The lot you receive is yours to inspect.
• Real-time stock or pricing. Our scores do not update on every promotion; check the vendor directly.

Methodology Changelog

Changes to the formula, the weights, the gate or the lab rubric are recorded here with their date.

• 2026-11-06: Shared analyses and lot differences (A7). The same laboratory analysis issued to two vendors now has a stated rule: no deduction when the report names the vendor as client, and the match is disclosed on both reviews. LOT also covers a posted certificate printing a lot that differs from the laboratory's record of that report.
• 2026-10-29: Rule A4, a certificate that reports its own failure.
  • A certificate now scores 0 when the document itself reports a failed identity, purity or specification result, as well as when the laboratory's own record does. Until now a failure printed on a certificate whose key the laboratory did not find was not covered, because the FAIL case rested on the laboratory's record. Applied first to AminoWave Labs, reviewed on 27 October 2026. It applies to every vendor; no other listed certificate reports a failed result.
• 2026-10-27: Rule A6, which certificate counts when there are several.
  • When a vendor posts more than one certificate for a product, across sizes or lots, the product is now scored from the newest certificate by its A6 date. The rules did not say which certificate to use before. Applied first to StudzPeptides, reviewed on 27 October 2026. Vendors listed earlier are brought in line at their next re-audit, and any score change is recorded in their audit history.
• 2026-10-25: Pricing rule E2, products sold only in packs.
  • New rule E2: when a vendor sells a size only as a pack of several vials, and no single vial is sold, the price is the pack price at the minimum order, and the price per mg divides it by all the milligrams in the pack. Discounts that need more than the minimum order, such as a lower price for several packs, are not counted. The vendor page says that the vendor sells only in packs. Until now, multi-vial packs were kept out of the comparison, which left a vendor that sells only packs with no comparable prices.
  • Applied to Peptide Warehouse US, first reviewed on 25 October 2026, which sells every product only as a 10-vial pack with a minimum order of one box, and offers 5%, 8% and 12% off at 3, 5 and 10 or more boxes. Each pack price is divided by its 10 vials; the box discounts are not counted. No other listed vendor sells only in packs, so no other vendor's own prices changed.
  • Pack-only prices are compared against the market, but they are not included in the market medians used for other vendors, because they are minimum-order bulk prices (owner ruling of 26 October 2026, the day after the rule was first drafted). Adding Peptide Warehouse US therefore changed no other vendor's pricing sub-score or score, and no testing-gate result changed.
• 2026-10-24: Pricing rule E, price at a vendor's first review, and a cap on free-vial offers (E1).
  • Rule E now states how a sale is treated at a vendor's first review: the price is the one a buyer pays that day, unless the sale has a stated end date or is tied to an event or season. The 30-day test applies to price changes we observe after the vendor is listed. For vendors listed earlier, the 30-day test still applies to the price changes seen since their first review.
  • Applied to Bionova Peptides, whose first review was on 23 October 2026: every product was shown at half its struck-through regular price, with no stated end date and not tied to an event or season, and the same prices were still live when we re-read them on 24 October. Its listings now carry the prices a buyer pays instead of the regular prices. Its pricing sub-score moved from 2.4 to 8.1 and its score from 5.19 (Weak) to 6.05 (Fair); no testing-gate result changed.
  • Bionova Peptides' prices also lower the market median for the same products, so pricing was recomputed for every vendor: 25 other vendors' pricing sub-scores fell by 0.1 to 0.3 and their scores by 0.01 to 0.05. One vendor moved from Strong to Fair. No testing-gate result changed. Each vendor's audit history records its change.
  • Rule E1 now caps free-vial offers: at most 2 vials per purchase are counted, and only when a single vial is also sold on its own; an offer of more than 2 vials for one total is not counted. The one offer counted under E1, Zeptix Labs' buy-one-get-one-free, gives 2 vials and sells a single vial on its own, so no score changed.
• 2026-10-23: Pricing rule E1, standing offers and free vials (wording revised the same day).
  • First published and revised on the same day. The first version excluded offers where a single vial is also sold at the same price. The revision counts the lowest total a buyer can pay under standing offers: for each size, the price is the lowest total a buyer can pay without a code under the vendor's standing prices, and when that purchase includes more than one vial, the price per mg divides that total by all the milligrams received. An offer counts as standing when it has no stated end date, is not tied to an event or season, and was live at two checks at least 30 days apart. Offers that need a code, time-limited sales, and offers that require spending more than the single-vial price are not counted.
  • Each counted offer is recorded with the vendor's wording, the page, the date it was first seen and the date it was last checked. An automatic check re-reads the vendor's page every day and a second check reviews each offer every month. A size whose offer has not been confirmed live for more than 35 days, or that shows an end date or an event, is priced at its single-vial price again.
  • Under the first wording no offer on record qualified, so no price was counted under E1 and pricing was recomputed for every vendor from current listings with the published method. Listings added or updated since 22 October, including the listings of a newly reviewed vendor, moved the market median for the same products, so 53 vendors' price ratios moved slightly and 26 pricing sub-scores changed by 0.1 to 0.3. 26 vendors' scores changed by 0.01 to 0.05; no vendor changed tier and no testing-gate result changed.
  • Under the revised wording one offer qualifies: Zeptix Labs' buy-one-get-one-free, which needs no code, states no end date, is not tied to an event and has been live since early September 2026, and under which a single vial costs the same as the pair. 8 of the 10 sizes on the offer count now; the other 2 count once they have been live for 30 days. The other free-vial offers on record do not qualify: one is tied to Halloween, one gives the second vial at half price and ends on 31 October, one needs a code, and one was first seen on 21 October.
  • Pricing was recomputed again with the revised wording. Zeptix Labs' pricing sub-score moved from 6.9 to 8.8 and its score from 7.89 to 8.17. Its offer also lowers the market median for the same products, so 14 other vendors' pricing sub-scores fell by 0.1 to 0.2 and their scores by 0.01 to 0.03. No vendor changed tier and no testing-gate result changed. Each vendor's audit history records its change.
  • Prices re-checked the same day against the rule above. For three vendors, prices had been taken from a time-limited sale (one with a stated end date or tied to an event), and they were corrected to the vendor's standing prices. For one vendor, an undated sale price charged without a code replaced the regular price recorded before, and one vendor's prices were updated after its regular prices fell. The pricing bullet on standing sales was reworded to match E1. Pricing was recomputed: 13 vendors' scores changed by 0.01 to 0.20, one vendor moved from Fair to Weak because lower prices for the same products moved the market median, and no testing-gate result changed. Each vendor's audit history records its change.
• 2026-10-22: Laboratory rules tightened after a review of the rulebook.
  • Tier C is now decided from the check results alone: L1 and L2 pass, L3 fails, and neither L4 nor L5 passes; otherwise a failed L3 gives B-. It replaces the words "little else documented", which left room for judgement. 1 laboratory moved from B- to C (Analiza Białek).
  • Every laboratory is graded (L0): a laboratory is graded as soon as it is first named in our records, on a certificate or in a vendor's claims, from its own public reports for any client. The 4 laboratories previously listed as not graded now have tiers (Liquilabs B+, APIGENEX B-, Colmaric Analyticals C, SafeCert Labs D). None of them is named on a certificate we score, so no vendor score changed.
  • ChemClarity was removed from the lab directory: its website describes a nonprofit chemical-information initiative, not a testing laboratory, and no current vendor or certificate names it.
  • New L1 clarification (linked registry entry). Applied to BioRegen, whose L1 now passes on the Wyoming and Texas registrations of BioRegen, LLC, linked through a named chemist's public profile and a job posting; its tier follows from its checks (A). No score changed.
  • L2c (shared ownership) rewritten: a link to a listed vendor is acceptable only when the laboratory serves unrelated clients, passes L1 to L4 and passes a documented deep check; a proven link that fails any of these grades the laboratory D. No laboratory currently has a proven link, so no score changed.
  • The accreditation point now applies only to certificates tested on or after the date from which the accrediting body's register shows the accreditation as active. MDx BioAnalytical Laboratory's PJLA accreditation is active from 9 June 2026; 24 of its certificates were tested earlier and no longer earn the point. 2 vendors' scores changed.
  • A certificate's laboratory name must now match a laboratory's name or a recorded spelling exactly; a partial match no longer counts. Recorded spellings were added for 6 laboratories. No score changed.
  • L5 is applied as written, to reports for any client: 4 laboratories were re-checked (ACS Laboratory, Accurate Test Lab, Harrens Lab and Vanguard Laboratory) and all 4 now pass L5; ACS Laboratory moved from A- to A. No score changed. The page now states that tiers B- to A+ earn the same laboratory points.
  • Laboratory profiles corrected: Freedom Diagnostics Testing's L1 record (still a pass) and BioChex's L1 record (still a fail) now state what was checked.
  • Effect: 3 laboratories changed tier among those named on scored certificates (Analiza Białek B- to C, ACS Laboratory A- to A, BioRegen to A from an undecided A-), 4 were graded for the first time and 1 was removed. 3 vendors' scores changed because of these rule changes (Amino Amigos, AminoVault, BioLongevity Labs), and 2 more on the same day because of confirmed certificate findings (Biotech Peptides, Raw Amino). No testing-gate result changed. Each vendor's audit history records its change.
• 2026-10-22: Laboratory rulebook (L-rules). Every laboratory in the directory is now graded by six published checks (L1 to L6) and a fixed tier table, replacing the earlier descriptive tiers ("Named + Verification Portal + Full Panel" and similar), which left room for judgement. Each laboratory's profile now shows the result of every check, its evidence link and the date it was checked. A laboratory from which we hold no certificate is listed as not graded (L0) and has no effect on any score. The +1 accreditation point now requires an active ISO/IEC 17025 accreditation on the accrediting body's own register with a scope that covers peptide testing (L6); a laboratory's own claim, or an accreditation for other work, no longer earns it. The certificate rubric itself (A2, A3, A9) is unchanged; laboratories changed tier as follows: 18 laboratories regraded (17 down, 1 up), 4 lost the accreditation point, 5 listed as not graded. Effect: 9 vendors' lab sub-scores changed (8 final scores fell, one was held by the testing-gate cap), 0 testing-gate results changed, 2 vendors changed score tier. Each vendor's audit history records its change.
• 2026-10-22 (later the same day): Core and niche products (A1b). A missing certificate on one of the 15 most widely carried products still counts fully; on any other product it now counts at half weight, so a broad catalogue is not penalised for its breadth as much as before. 33 vendors' lab scores rose; none fell.
• 2026-10-22: Rules version 2. Lab scoring changed today, and every rule is now printed on this page.
  • Correction first: earlier on 22 October we applied one lab rule to every vendor (fixed points per certificate and fixed deductions), replacing lab values that had been held from earlier audits, and the entry below said lab scoring was unchanged. That was wrong; lab scores did change that morning, and each change is in the vendor's audit history.
  • A certificate that names no laboratory now scores the flat 2.0, the same as a self-issued certificate or one from a laboratory whose independence is not established (A3). Before, it could score 3 or more.
  • A certificate now scores 0 when the laboratory's own record reports a FAIL, or when the vendor's posted copy differs materially from the laboratory's copy (A4). This replaces the COPY deduction (−0.2 each, cap −0.4), which was taken off the whole lab sub-score.
  • A key printed but not found at the laboratory keeps its +1.5 for the printed key and earns nothing for confirmation (A5). The CODE deduction (−0.1 each, cap −0.3) is retired, and codes on certificates we do not score no longer count against a vendor.
  • Undated certificates are aged from the earliest date we can prove they existed, and count as older than 12 months when there is none (A6). Before, an undated certificate never aged.
  • Microbial-limits screens (TAMC, TYMC, USP <61> and <62>) no longer earn the sterility points, for every vendor. A report testing only endotoxin, sterility or heavy metals is never scored as a product's certificate (A2b).
  • The only adjustments left are LOT, SUPPLIER and CLAIM, with the amounts and caps printed above; CLAIM no longer applies to products that simply lack a certificate.
  • The testing gate is now computed from the certificates (A9) instead of being set by hand: at least half of a vendor's products need a lab-confirmed certificate from a directory laboratory that is not tier D.
  • The +0.5 for a state-registry record in transparency is removed for every vendor (C2); transparency now tops out at 9.0.
  • A store not taking orders keeps its sub-scores and gate from its last full audit (F1). One store that had been re-scored from its closed pages was restored to its 26 September 2026 audit; it later closed permanently and is no longer listed.
  • Every rule, point value, the reason for each, the gate threshold, rounding, ties and product-page order are now printed here, and the site's scoring code is tested against this page before every update.
  • Effect: 22 vendors' scores changed and 26 testing-gate results changed (25 to "not verified", 1 to "verified"); 3 vendors changed tier. Each vendor's audit history records its change and the rule behind it. The weights, the tier thresholds, the gate cap and the shipping, pricing and operating-history point tables are unchanged.
• 2026-10-22: One published rubric for shipping, pricing, operating history and transparency. Every vendor was re-scored on the same day under the point tables printed above:
  • Prices were re-read from every vendor's site and matched to the exact milligrams per vial; a standing sale price counts as the price, and the deduction for struck-through prices is withdrawn.
  • Operating history now measures the business rather than the domain, from the earliest dated evidence that it was operating.
  • Shipping and transparency are scored item by item from the vendor's own published policies, with each address checked as physical, virtual or mailbox.
  • Research content on a vendor's site is no longer scored. A temporary outage, such as card payments being down, is noted rather than scored.
• 2026-10-21: Methodology page brought in line with the scoring in use. These are clarifications of how scores are calculated, published together:
  • The coverage rule is applied to every vendor: every catalogue product a vendor sells counts toward the lab sub-score, and a product with no published certificate counts as 0.
  • A certificate for another size of the same product counts for that product; a blend certificate counts only for that blend.
  • Lab supplies such as bacteriostatic water are excluded from the lab average.
  • The lab sub-score floor is 0, not 1.0: a vendor with no published certificates scores 0.0 on lab.
  • The per-certificate lab rubric, the testing-gate cap (maximum 8.5), the tier thresholds (Excellent 8.5, Strong 7, Fair 5.5, Weak 4) and the current evidence tiers are published in full on this page. The star scale is retired; scores are shown out of 10 with a tier label.
• 2026-05-20: Weights rebalanced: lab 50→55%, operating history 10→5%. Lab Credibility went from 50% to 55% (the most defensible signal in this market) and Operating History from 10% to 5%. The vendor market changes quickly, and a heavier tenure weight was holding back newer vendors with verifiable testing. Every vendor was re-scored under the new weights.
• 2026-05-19: Score-integrity check and purity-adjusted lab score. Compared every vendor's stored lab sub-score with the average of its carried products' lab scores and corrected the one that did not match (BioLongevity Labs, lab 10 → 4.21, where 6 of 7 products named a laboratory with no certificate published). Also introduced the purity bands for every product with a measured purity figure.
• 2026-05-19: Evidence tiers and the verification page. Every carried product was placed in an evidence tier, and a new verification page began publishing the per-vendor evidence breakdown. The tiers have since been revised to the five described above (lab-confirmed, read and checked, linked, claimed, none).
• 2026-05-18: Directory expansion (16 → 49 vendors). Added 33 peptide vendors based in the US, scored on the same five-dimension formula and anchor scale. One candidate (Limitless Life Nootropics) was identified as the same business as an existing entry (Limitless Biotech, both Live Limitless LLC, Gulf Breeze FL) and was not added twice.
• 2026-05-18: Anchor recalibration (lab and transparency). The first scoring anchors were too low, and lab and transparency sub-scores were raised by a fixed amount for the vendors then rated. That adjustment no longer affects any score: every lab sub-score is now computed from the rubric above, and every transparency sub-score from the point table above.