How PeptideBase works.

PeptideBase is an independent index of the peptide ecosystem. We collect what public sources say about providers, research vendors and compounds, structure it, and show the source for every claim.

We are not a clinician. This page exists so you can judge exactly how much weight our data deserves.

PeptideBase is

• An independent, editorially neutral directory and research index
• A structured starting point for your own research
• A record of what public, checkable sources say — each one named

PeptideBase is not

• A medical service, telehealth platform, pharmacy or prescriber
• A source of dosing guidance, protocols or treatment recommendations
• A quality guarantee, safety certification or endorsement
• A replacement for a licensed clinician who knows your history

On inclusion and absence

A provider appearing on PeptideBase is not endorsed by PeptideBase. A provider not appearing has not been rejected — in almost every case it means we have not indexed it yet. Absence is a gap in our coverage, not a verdict.

We hold three kinds of record. They are assessed three different ways, because three different amounts of checkable evidence exist for each.

We order providers. We do not score research vendors. We do not grade compounds. Each is a deliberate choice, explained below.

What we don’t draw on, and why

We do not use FDA import alert records. The public listings redact firm names, so the data cannot be matched to a provider. We tried; it yields nothing.

We have not completed a systematic search of FDA warning and enforcement letters. Where we hold a warning-letter finding it has been checked by hand, and where a regulatory concern is material it appears on the provider’s record. Absence of a warning on a listing means we have not searched — not that none exists.

Regulatory records show that an action exists. They do not show its outcome or its merit, and we do not independently assess whether a concern is valid.

The rules we hold ourselves to

• We do not invent data. Where we hold no source for a field, the field is empty. We do not estimate or infer.
• Every rating carries its source. Where we hold no source, no rating is shown.
• We do not adjust ratings. No normalising, weighting or recalculating. What the third-party platform reported is what you see.

Providers are ordered by a listing score out of 100, computed in our database and recalculated whenever a listing changes.

Here is what the score actually rewards:

Every row above is a completeness or reputation signal. Until 14 August 2026 there was one that was neither: a 5-point slot for featured status, with the directory’s sort going further and pinning a featured listing above the score entirely. Both are gone — the slot from the formula, the pin from every query that ordered listings. Neither had ever been used. We disclosed them while they existed rather than publish a tidied formula, and we are recording their removal here for the same reason.

What this score does in practice

Manual review dominates everything else. Listings we have reviewed average 82 out of 100. Listings we have not average 15. (Measured across all listings in our database, 13 August 2026.) The remaining components separate listings within those two groups, but rarely move a listing between them.

So the honest description of our default order is: listings we have reviewed, above listings we have not. The rest is refinement.

What this score does not measure

Nothing in it reflects clinical quality, safety, patient outcomes, price, or whether a licence is currently valid. More than half the available points are awarded for having filled fields in. Fewer than half our listings carry a rating at all, so for most providers that component contributes nothing.

A higher score means a more complete, more recently reviewed listing. It does not mean better care.

What we are changing

We are not satisfied with this. A score this dependent on one manual flag measures our own coverage more than it measures the provider. We are rebuilding it. When we do, this section changes and the changelog records the date.

This is not a ranking. We don’t score vendors or tell you who’s best — we show you what each one makes public and let you decide.

We check six signals:

1. An openable laboratory report
2. A testing claim
3. A named testing laboratory
4. Shipping information
5. A contact address
6. A public X handle

Only the first one is proof — it’s a document we can open and read ourselves. The other five are the vendor telling us about themselves.

Vendors are grouped by what that check found:

“Nothing at all” means we found no public documentation. It does not mean a vendor’s product is untested — only that nothing was published where we could find it.

For each compound we show indexed citations, regulatory context, provider availability, and — where our thresholds are met — a median price per mg.

Indexed citation counts measure how much research we have indexed, not how strong that research is. A higher count is not an endorsement.

When we show a median price, and when we don’t. We publish a median only where at least five distinct providers list a price for that compound, and where the spread between the 25th and 75th percentile is wide enough to be meaningful. Prices below an extraction-confidence threshold, outside a plausibility range, or older than 45 days are excluded before the median is computed. Where those conditions aren’t met we show no median rather than a thin one.

Currently on hold. Our dedicated price pages are temporarily showing a maintenance notice while we rework the pricing pipeline.

We do not grade compounds, and we do not publish dosing guidance, administration routes or protocols — not as an oversight, but because doing so would make this a clinical tool, which it is not.

Much of the peptide literature is preclinical or conducted in narrow populations. A compound with substantial indexed research may still have no evidence base for the use you have in mind.

Our matching tool is offline. It is being rebuilt and is not currently available.

When it returns we will document how it works here — including what it screens for and what it does not — before it goes live, not after.

We withdrew the Verified badge in July 2026. It is no longer displayed anywhere on PeptideBase.

The reason is straightforward. The label promised more than the check behind it delivered. What it actually confirmed was that a website was reachable and a business was classified. It did not confirm business registration, licensing, or product quality — which is what most people reasonably assume a Verified badge on a health-adjacent directory means.

A badge on a directory like this should mean something specific. That one did not.

The absence of a badge has never meant a provider failed a review, and it does not now.

Nothing has replaced it. No provider on PeptideBase carries a trust mark of any kind, and we will not introduce one until we can publish the exact check behind it. We have retired the word itself as well — it previously appeared as a directory filter and a headline count, and both implied a verification step we do not perform.

One thing we want to be plain about. The same manual review that sat behind that badge is still the largest single component of how the directory is ordered — the 30 points described in §04. We think disclosing that is better than quietly keeping it, and better than pretending the two are unrelated. It is also the main reason we are rebuilding the score.

What this methodology cannot tell you. We would rather you knew before relying on anything here.

1. We assess what is publicly documented. A provider that documents little may be excellent; a provider that documents well may not be. Our method cannot tell the difference.

2. We do not inspect, audit or test. We do not visit facilities, review records, test products, or confirm that a licence is valid on the day you read this page.

3. Ratings measure reporting, not care. We do not assess whether individual reviews are genuine, and review volume tracks a provider’s age and marketing budget at least as closely as its quality.

4. Regulatory records show presence, not outcome. An indexed action means an action exists. It does not mean wrongdoing occurred, and we do not adjudicate it.

5. Our coverage is uneven. A provider’s absence almost always reflects the limits of our coverage, not a judgement about them.

6. Our ordering reflects our own work as much as the provider’s. The largest component of a listing’s score is whether we have reviewed it. A provider we have not got to yet ranks low for that reason alone.

7. The underlying science is early. Much peptide research is preclinical or conducted in narrow populations. Volume of literature is not strength of evidence, and we do not claim to distinguish between them.

Reporting an error

Email [email protected] with “Correction” in the subject line. Tell us the record, the specific field, and the correct value. Include a public source where you can.

What we do

We verify against the source, correct the record, and update it. Corrections that change a score, a documentation state or a regulatory signal are logged in the changelog.

Claiming or updating a listing

Providers can update a listing by emailing [email protected] with “Provider listing” in the subject line. We verify submitted information against public sources before publishing it. Being the subject of a listing does not give you editorial control over it.

Disputing an assessment

Tell us the specific field or score you dispute, why, and include supporting documentation. We respond within 14 days.

Removal

Any listed business may request removal, and we honour removal requests. We do not remove or suppress accurate information in exchange for payment, advertising, or any other consideration.

How we make money

We don’t. PeptideBase currently generates no revenue — no advertising, no sponsorship, no affiliate commission, no paid listings, no fees of any kind.

If that changes, it will not change in one direction: we will not take money from the companies we list. Not for placement, not for ranking, not for a badge, not for a sponsored slot, not as commission on something you buy from them, and not to claim or correct a listing. That rules out most of the ways a directory in this category gets funded. It leaves other routes open — reader-side subscriptions, licensing our data to people who are not listed in it — and if we ever take one, this page changes before the money does.

Running unfunded is a constraint, not a virtue. It limits how quickly we can review listings, and our limitations reflect that directly: the largest component of a listing’s score is whether we have got to it yet.

No paid placement

No provider pays to appear or to rank. Zero listings in our database carry featured, pinned or sponsored status — not one, out of 3,822 — and this page re-verifies that count every time it is built. No provider has ever been placed above another for money.

The structural reasons you can check that

1. No commercial relationship can contribute a point to any listing’s score — not merely that none has. The formula contains no commercial component to contribute one, and no query pins a listing ahead of its score. A 5-point slot for featured status existed until 14 August 2026 and is recorded in §04 and the changelog. You can check this one by reading the formula, rather than by trusting a statement about the past.
2. Our build fails if any listing is given featured, pinned or sponsored status while this page says otherwise, and an independent check re-runs the same assertion against our production database daily.
3. Automated checks block our deploys if affiliate or commission language appears anywhere in our copy or our published content.

Funding

No venture funding. No acquisition pipeline.

Who runs this

Founded by Alexandre Peterson, run by a deliberately small independent team.

On our credentials — stated plainly

PeptideBase is not currently staffed by licensed clinicians. We are recruiting credentialed reviewers — physicians, pharmacists and researchers — to review our educational content. Until that review is in place, treat everything on this site as preparation for a conversation with your own clinician, not a substitute for one.

Our versioning policy

Superseded methodology remains available at its dated URL. We relabel retired methodology; we do not delete it.