How an answer earns its grade.
Trust in a medical answer shouldn't rest on tone. Here's where the evidence comes from, how it's graded and when the engine says no.
Sourcing
Answers draw on the public medical literature and official drug labels: PubMed, PubMed Central, ClinicalTrials.gov, DailyMed and openFDA. Marketing pages and unsourced blogs never enter the pool.
Grading
Each answer is graded A to D on the strength of the studies behind it: their design, how many people they studied and whether they agree. One small trial doesn't read as settled science.
Citation checks
Before an answer goes out, every claim is checked against the citation attached to it. A statement that can't be tied to a real source span doesn't ship.
Refusal
When the sources can't support a safe answer, the engine refuses and says why. An interaction the labels don't describe is reported as no data, never as safe.
What a grade means
A High
Well-run trials that agree with each other. You can lean on it.
B Moderate
Good studies with limits in size or design. Read them before you act.
C Low
Small, early or mixed studies. A reason to look closer, not a plan.
D Very low
Thin or conflicting. Treat it as a hint and nothing more.
Evidence too thin
The literature can’t support an answer, so the engine refuses and gives the reason.
Refused
A safety rule fired. A personal dose for a compound with no label dosing is never given, whatever the studies say.
One question, start to finish
A real question, the grade it earned and the studies it stands on. Every answer has this anatomy: a claim, a grade and sources you can open yourself.
Question
Does creatine supplementation harm kidney function in healthy adults?
Grade B: Moderate certainty
In healthy adults, controlled studies haven't shown that creatine at common doses damages kidney function, though most trials are short and exclude people with existing kidney disease. Serum creatinine can rise on creatine without reflecting real injury, which complicates routine monitoring. For anyone with reduced kidney function the evidence is too thin to reassure.
- PMID 28615996Kreider RB et al, J Int Soc Sports Nutr 2017
- PMID 18188581Gualano B et al, Eur J Appl Physiol 2008
- PMID 33557850Antonio J et al, J Int Soc Sports Nutr 2021
An illustrative answer in the engine's format, condensed for this page. The grade and the sources are the ones it rests on, and every PMID opens the record it names.
What it won't answer
Being honest means naming the questions the engine hands back to you. These stay outside its lane on purpose.
Diagnosis and treatment decisions
The engine informs your judgment. It doesn't make the call.
Dosing
Generated protocols withhold dosing. An answer can quote the dose printed on an official FDA drug label, with the label cited. A dose is never invented.
Questions with no literature
When the evidence is absent, the engine refuses instead of guessing.
Anything that needs the patient's record
It reasons from published evidence, not from chart data it doesn't hold.
Questions
Is this medical advice?
No. It's a research tool for licensed professionals. It supports your judgment and points you to sources. It doesn't replace either.
Where do the answers come from?
The public medical literature and official drug labels: PubMed, ClinicalTrials.gov and DailyMed among them. Every claim links back to its source.
What do the grades mean?
Each answer is graded A to D, from high certainty to very low, on the design, size and consistency of the studies behind it. Weak evidence is labelled weak.
What happens when there's no good evidence?
The engine refuses. It tells you the literature can't support a confident answer instead of inventing one.
Does it check drug and supplement interactions?
Yes. Enter a stack and each pair comes back with its label-stated interaction and citation. Compounds it doesn't track get flagged, not passed.
Can I build on it?
Yes. The same engine is available over an API with grounded responses, safety endpoints and tiered keys.
Run the method on a question you can check.
The free tier is the same engine. Ask it something you already know the answer to and read the sources it picks.