Editorial standards
How we research, grade and write
These are the rules every analysis on this site is held to. They exist so a reader can judge our work rather than take it on trust.
Last reviewed: September 2026
Primary sources, named and linked
Every claim about what research shows is tied to a specific paper with authors, year, journal and a direct DOI or publisher link, so a reader can open it and disagree with us. We do not cite other websites' summaries of a study as if they were the study.
No government sources
By editorial policy this site draws on peer-reviewed literature and named independent sources rather than government or regulatory publications. Where a legal or regulatory question matters to a reader's decision, we say that it does and direct them to qualified advice instead of paraphrasing policy.
Therapeutic research is not use research
Most human data on these compounds comes from patients at medical doses under supervision. Recreational use is typically higher-dose, unmonitored, multi-compound and of uncertain product quality. We keep that distinction explicit rather than presenting trial safety data as reassurance.
Community reports stay separate
Reported experience is genuinely informative about what people actually encounter, and it appears in a clearly marked block that never sits inside the cited-research section. We do not invent reports, quotes or forum posts, and where no verifiable community feedback exists we say so rather than filling the space.
Uncertainty is stated, not smoothed
Every analysis carries an explicit unknowns section. If the honest answer to a question is that nobody has studied it in humans, that is what the page says.
No dosing, cycles or protocols
We describe what studies administered when reporting their findings. We do not translate that into instructions, ranges or recommendations, and we do not publish cycle or recovery protocols.
Corrections are public
Errors are fixed on the page with the change stated, not edited away quietly. Anyone can challenge a claim through our corrections process.
The evidence-grading scale
Every compound analysis carries one of four grades. The grade rates the strength of the published literature — not how safe, effective or advisable a compound is.
- Moderate
- Multiple human studies, including controlled trials, with reasonably consistent findings. Still usually conducted in clinical populations at therapeutic doses, so it may not transfer to recreational use.
- Limited
- Some human data exists but is small, short, uncontrolled, or confined to one population. Enough to describe a direction of effect, not enough to quantify risk.
- Very limited
- Isolated human studies, case reports or forensic literature only. Often the only human evidence is what went wrong.
- Preclinical only
- Cell or animal research with no completed human trials. A mechanism worth studying — not evidence that something works or is safe in people.
Related
Vendor scoring is documented separately on the assessment methodology page. Unfamiliar terminology is defined in the glossary. To challenge something we have published, see corrections and disputes.