How a page here is made
Most health sites ask you to trust the output. This page describes the process instead, so you can judge it — including the parts that should make you cautious.
#Who writes it
Pages are drafted with AI assistance and edited by a human. They are not reviewed by a clinician. That sentence appears at the foot of every protocol, and it is meant literally: no doctor, physiotherapist or dietitian has signed off on this content. If a page ever claims a professional reviewer, it is stating a fact about a specific named person — and no page does, today.
Nothing here is written for a sponsor. Nothing is for sale, there are no affiliate links, and no commercial product or supplement brand is ever named as something to buy. Independent testing standards (NSF, USP, Informed Sport, IFOS) are named, because "look for third-party testing" is advice you can act on without being sold to.
#Where the words come from
The site is generated, not hand-maintained page by page. A build step reads the source — markdown for the compound corpus, structured data files for causes, pathways, molecular targets, exercises and protocol plans — and writes every page from it. The consequence worth knowing: a correction made once propagates everywhere that claim appears, and an error made once does the same. That is why the automated checks below exist.
#How evidence is rated
The 1–5 star rating summarises human evidence for a compound overall — not for your particular goal, and not how well it will work for you. Animal-only evidence is capped at two stars and has to say "animal" on the page. This is enforced at build time rather than by editorial discipline, because it had previously failed: one compound was showing five stars off rodent data. The full scale is on the legend page →
Where a page cites a trial, the citation is bound to a real record and links out to it — usually PubMed or Europe PMC, where a large share of the papers are free to read in full. A citation here is a pointer you are meant to follow, not a decoration.
#What we will not publish
- No doses for prescription medicines. Prescription-only and controlled medicines are documented so you know they exist and can raise them with a clinician. They are listed, not ranked, and never recommended.
- No head-to-head "which is better" comparison involving a prescription medicine. Ranking a medicine you cannot buy against a supplement you can is not a useful comparison, and in Singapore advertising a prescription-only medicine to the public is prohibited outright — the Medicines Act 1975 and its advertising regulations carry no educational exemption.
- No diagnosis. The root-cause questions narrow your reading. They are not a clinical assessment.
- No certainty we do not have. Where the honest answer is "no supplement has trial evidence for this", the page says that instead of filling the space.
#The checks that can stop a release
Editorial rules that live only in someone's head get rediscovered as bugs. So the rules that matter most are executable, and a failure blocks publication rather than filing a ticket. The build currently refuses to ship if:
- an animal-only compound shows more than two stars, or carries a rating badge that does not say "animal";
- a page names a prescription or controlled substance without stating that it is prescription or controlled;
- a restricted compound's page would render self-dosing instructions;
- a dose calculator lacks a machine-readable maximum (a milligram/gram slip once made one calculator understate a dose a thousandfold);
- a protocol would prescribe a movement that the same page tells you to avoid;
- an evidence claim is not bound to a record in the claims file;
- any compound is missing from the goal taxonomy;
- a machine-readable data block would be emitted in a form search engines silently discard;
- a page links to a route that does not exist, or a page is published with nothing linking to it.
Each of these exists because that exact defect shipped at least once.
#The limits you should hold in mind
- No clinician review. This is the largest one.
- A star rating is a summary of a literature, and summaries lose information. The page body is where the honest detail is; read it before acting on a number.
- Coverage is uneven. Some compounds have a deep evidence layer, others a short profile.
- Singapore availability and regulatory status change, and this site is a snapshot.
#Found something wrong?
That is the most useful thing you can send. Errors here should be fixed rather than defended. See what has already been corrected, and how to report one →