Standards
How we decide what counts.
A database is only as useful as the rules behind it. Ours are published in full, in plain language, so you can check our work rather than take it on trust.
The A–F evidence rubric
Every compound carries one grade. It answers a single question: how strong is the published evidence that this compound does what it is claimed to do in humans? It is not a rating of safety, popularity, potency or value.
Multiple human RCTs
Two or more adequately powered randomized controlled trials in humans, with consistent results and published outcome data. Regulatory approval is common at this level but is not itself the criterion.
Human trials, limited scope
At least one well-conducted randomized human trial, but the evidence base is limited — a single trial, phase 2 only, small samples, or surrogate rather than clinical endpoints.
Preclinical or uncontrolled
Consistent preclinical signal with no adequately powered human trial, or human data limited to uncontrolled and observational studies. Findings may be concentrated in a single research group.
Weak, conflicting or negative
Human evidence is minimal, contradictory, or the best-designed trials failed to separate from placebo. Also assigned where a real evidence base exists but has never been independently replicated.
No credible evidence or net harm
No credible evidence of the claimed effect in any species, or well-conducted human evidence showing harm that outweighs any documented benefit.
What a grade is not
- A safety rating. Grade A means well studied, not well tolerated.
- A recommendation. Nothing here suggests any compound should be used.
- A permanent judgment. Grades move when the literature moves.
- A measure of interest. The most discussed compounds are often the least studied.
Citation sourcing policy
Primary sources only. Every claim on a profile traces to one of three places.
PubMed-indexed literature
Peer-reviewed papers, cited with a PMID so the source is one click away and independently verifiable.
ClinicalTrials.gov registrations
Trial design, phase, enrolment and status taken from the registry entry rather than a press release about it.
Approved regulatory labelling
For approved compounds, dosing and contraindications come from the label itself.
What we never do
We do not copy or paraphrase other database sites, quote supplier marketing as if it were evidence, cite forum consensus, or reproduce a dosing figure whose origin we cannot identify. If a widely repeated number has no traceable source, it does not appear here — its absence is the finding.
How dosing is presented
Dosing tables record protocols exactly as their source reports them, with the source named on every row. They are descriptions of what was studied, framed as such, and never as guidance.
Where a figure comes from an animal model, the profile says so. Allometric scaling from a rodent dose to a human dose is not validated for most compounds in this database, so we do not perform that conversion or imply one.
Update cadence
Every profile shows a last-reviewed date, and it means what it says: a person re-checked the sources on that date. Profiles are revisited when a trial reports, when regulatory status changes, and on a rolling schedule regardless of whether anything has changed.
- 109
- Profiles maintained
- 335
- Primary citations
- 2026-07-28
- Most recent revision
Supplier vetting criteria
Evidence quality and product quality are separate problems. A supplier must clear all six criteria below — not a majority — to appear in the directory.
Third-party testing
Purity and identity confirmed by an independent laboratory the supplier does not own or control. In-house testing alone does not qualify.
Certificate of analysis on request
A batch-specific COA must be available to any customer for any lot, not a single representative certificate reused across production runs.
Batch traceability
Every vial carries a lot number that resolves to a specific COA. Without traceability, a certificate proves nothing about what arrived.
Accurate labelling
Products are labelled for research use with correct compound names and quantities. Suppliers making therapeutic claims are excluded regardless of product quality.
Track record
At least twelve months of operating history with consistent independent test results and no pattern of unresolved fulfilment complaints.
Commercial independence
Listing is never sold. An affiliate relationship may exist after a supplier qualifies, but it can never be the reason one qualifies, and it never affects an evidence grade.
Corrections
If a citation is wrong, a grade is indefensible, or a figure cannot be traced to its stated source, we want to know. Corrections are made to the profile and the last-reviewed date is updated. We do not quietly edit around an error.
Questions about the method
Does a high evidence grade mean a compound is safe?
No. A grade describes how much is known and how well it was studied, not whether a compound is safe or appropriate for any person. A grade A compound with a serious documented adverse effect profile is still grade A — the evidence for that profile is exactly what earned it.
Why do some widely used compounds have low grades?
Popularity and evidence quality are independent. Several of the most discussed compounds in this space have no adequately powered human trial behind them. The grading rubric measures the literature, not the level of interest.
Can a supplier pay for a better grade?
No. Evidence grades are assigned from published literature and nothing else. Supplier relationships are a separate part of the site with separate criteria, and no commercial arrangement has ever changed a grade.
How often are profiles updated?
Every profile carries a last-reviewed date. Profiles are re-checked when a new trial reports, when a regulatory status changes, and on a rolling schedule regardless. If the evidence moves, the grade moves with it.
Medical disclaimer
This page is for informational and research purposes only. Nothing here constitutes medical advice. The compounds discussed are research chemicals. Consult a qualified clinician before starting any protocol.