Methodology

How we assess UK research peptide retailers

The PeptideClear Trust Index assesses every UK research-peptide retailer it can find across three separate baseline pillars: Testing Transparency, Business Transparency and Responsible Practice, shown separately rather than blended into one score. This page covers the first two in detail. Testing Transparency asks whether a buyer can independently verify what a batch actually is, not whether the compound is safe, effective, or worth buying: we do not laboratory-test peptides or audit labs directly. Responsible Practice's methodology is covered on the Trust Index page itself (how we assess brands). Beyond the free baseline, retailers can also commission the deeper, paid Independent Verification process (54 checks across 8 areas): a separate tier, not part of this methodology.

Testing Transparency: the six scored signals

• · +2 Per-batch CoA published (not a single example standing in for the whole catalogue)
• · +2 Open access (no email, login, or form gate)
• · +2 QR-verifiable directly to a named third-party lab's own portal
• · +1 CoA supplied with the order (real, but not checkable before purchase)
• · +1 Named third-party lab partner (an actual identifiable lab, not "in-house" or "independent testing")
• · +1 Dedicated CoA URL or subdomain
• · -2 Publicly rejects third-party CoA testing
• · -1 CoA gated behind an email request
• · -1 CoA available only "on request", no defined process
• · -1 A CoA claim we could not substantiate despite having full access to check
• · Auto-red UK Companies House trading entity confirmed dissolved, regardless of score

Score sets the tier: green at 5 or above ("CoA: strong"), amber from 1 to 4 ("CoA: partial"), red below 1 ("CoA: weak / closed"). The same six signals are applied identically to every retailer; nobody can move a tier by hand. We display this as a score out of 100 (green 70+, amber 30-69, red below 30) because the raw point range is narrow enough that two very different retailers can look artificially close together; the 0-100 figure is a rescale for readability, not a different or more precise measurement.

Where we couldn't confirm something, we say so, and it isn't treated as a negative finding. A signal we couldn't observe (a site blocking automated access, a page that wouldn't load) is recorded as unverified or unconfirmed, never quietly turned into "no" or folded into a lower score as if it were a confirmed failure. The 143 retailers we couldn't assess at all sit outside the scored index entirely, for the same reason: not being able to check something is a different claim from it failing a check.

Business Transparency: a free-tier check, not just a paid one

For every retailer we search UK Companies House and record one of three states: confirmed (a matching registration found), not established (we searched by exact name, exact phrase, and plausible variants, and found nothing, despite the site presenting as UK-based), or not yet checked (we haven't completed this search for this retailer yet). These are deliberately different claims: a retailer we haven't checked is not evidence of anything, and is shown that way. A retailer we searched for and could not find is a real finding.

A retailer with "not established" business transparency is capped below green regardless of its CoA score, the same mechanism as the conduct cap above: a finding can withhold a "strong" verdict, never force a worse one. This exists because our own research found a retailer with a perfect CoA score and no findable UK company anywhere on the register, despite claiming UK jurisdiction on its own terms of service. Good paperwork about testing says nothing about whether we know who is actually selling it.

Needs re-confirmation: not a seventh tier

A small number of retailers carry "in-depth review needed" instead of a tier. This applies when a specific finding from our research (a Companies House match, a claim that didn't hold up when checked, a naming or registration inconsistency) is serious enough to disclose but not yet certain enough to stand behind as a verdict. It overrides the displayed tier, never the underlying score: an unresolved finding can withhold a published verdict, it cannot lower one. It carries no rating and is excluded from any structured Review data we publish. Each retailer in this state shows the specific unresolved signal, and moves to a normal tier once that signal is confirmed one way or the other.

Retailers we could not assess at all

Some identified retailers are excluded from scoring entirely rather than scored red: a dead domain, a site blocked to automated access, or a confirmed closed business. Not being able to check something is a different fact from it failing a check, so these sit outside the index until they can genuinely be assessed.

Signals that feed the Responsible Practice assessment

Separately from Testing Transparency, we also record observable signals that inform Responsible Practice (full framework: how we assess brands):

• · Sells peptides in oral capsule or tablet form (oral dosing by design, despite research-only framing)
• · No research-use-only disclaimer published anywhere, or missing from individual product listings
• · Publishes benefit or health claims next to products despite research-only framing
• · Registered office is a known virtual-office or mail-forwarding address
• · No card acquirer accepted (crypto or bank-transfer only)
• · Registered at the same address as another retailer scored in this index, at a known formation-agent building (a bare virtual office alone is not flagged; two unrelated retailers sharing one specific building is)

A signal here can withhold a strong Testing Transparency verdict; it cannot promote a worse one. If a retailer's own conduct contradicts its research-only claim, specifically an active benefit claim or no research-use-only disclaimer published anywhere, the six-signal Testing Transparency result alone cannot carry a "strong" verdict, whatever the six-signal number comes to. That result is capped at amber, with the reason stated, while the same signal separately feeds the Responsible Practice status shown alongside it.

These describe whether a seller's conduct matches its own research-only claim, which is a different question to CoA transparency. Quietly folding them into the Testing Transparency score would silently change every score already published, so they are shown, and assessed, separately.

What we explicitly do not measure

• · Biological activity of the peptide (a CoA does not prove activity).
• · Sterility (a separate analytical test, rarely provided by research peptide retailers).
• · Dose accuracy (a separate analytical test).
• · Customer outcomes (we do not solicit or publish user health outcomes from research peptide use).

Refresh cadence: what happens monthly

Retailer entries are re-checked on a rolling basis, not all at once. Each entry carries its own last-verified date on the PeptideClear Trust Index, not a single site-wide "last updated" claim. Concretely, each month:

• · A rolling slice of the index is fully re-checked against the six CoA signals, responsible practice, and business transparency: link liveness first (automated), then a full re-verification for anything that's changed, prioritising green-tier and previously-flagged retailers.
• · Every retailer carrying "in-depth review needed" or a "not established" business transparency is re-attempted, since these are the states most likely to resolve one way or the other over time.
• · The MHRA update feed is checked against every named retailer in the index, not just read as general news. The feed itself is automated (it updates at least every 45 days and is pulled at every site build), but cross-referencing its contents against our own retailer list by name is still a manual step today, done monthly.
• · ASA rulings are checked for anything naming a retailer already in the index.
• · Any single-sourced or unresolved standing caveat gets a fresh attempt at resolution. The Janoshik accreditation question, live at time of writing, is the example: checked monthly until it's answered from a primary source, then either cleared or turned into a permanent flag.

This cadence is a manual editorial process today, not an automated pipeline. We're saying so plainly rather than implying more automation than exists.

Conflict of interest

PeptideClear takes no commission or referral fee from any retailer scored here, and no payment can buy a better baseline result. That has been true since 19 July 2026. Retailers can separately pay for the deeper Independent Verification process, which buys the work, not a result. Full position at how we are funded.

Related: PeptideClear Trust Index (live counts and per-retailer detail) · full UK retailer comparison · Independent Verification · editorial policy · how we are funded.