Where our information comes from

Supplement marketing is loud. Research is quieter and more mixed. This site tries to show the research as it is, including when it says a popular supplement does very little.

1. Study counts come straight from PubMed

PubMed is the US National Library of Medicine database of medical research, with more than 36 million papers. For each of the 100 supplements we run a fixed search and count three things: randomized controlled trials in humans, meta-analyses and systematic reviews, and Cochrane reviews. Together that is 55,051 papers across the site.

Every number is a link to the exact PubMed search behind it, so you can check it yourself. The counts include every matching study, for any use, whether it found a benefit or not. A big number means a supplement is well researched, not that it works.

2. Key reviews are real, linked papers

On each supplement page we list up to six reviews that name the supplement in their title, Cochrane reviews first. Cochrane is an independent network known for the most rigorous reviews in medicine. Each entry links to the paper on PubMed by its PMID, its permanent ID.

3. Reference fact sheets

For doses, upper limits and interactions we rely on government sources written by scientists without a product to sell: the NIH Office of Dietary Supplements, the National Center for Complementary and Integrative Health (NCCIH) and the FDA.

4. The dose that worked in studies

The dark dose box on a supplement page comes from the same engine the Underdosed app uses to check scanned labels: the daily range used in the research, the official upper limit, and which chemical forms your body absorbs well. Where an official upper limit exists, the page shows the government figure.

5. How we grade each claim

A supplement is never graded as a whole. Each claim gets its own grade, because creatine can be excellent for strength and unproven for memory at the same time. The grade reflects the best available reviews and trials, how consistent they are, how big the effect is, and who paid for the research.

Strong evidence. Consistent results across good human trials or meta-analyses.

Moderate evidence. Several decent human trials, mostly positive. The effect may be small.

Limited evidence. Few or small trials, mixed results, or mostly industry funded.

Weak or no evidence. Little good human evidence, or trials mostly negative.

When something works against a goal (caffeine and sleep, for example) it is shown as a warning, never as a benefit.

6. What we have not done yet

To be upfront: the summaries and grades were drafted with AI assistance from the sources above and have not yet been reviewed by a doctor or dietitian. We are working on clinical review. Until then, treat this site as a starting point for a conversation with a professional, not a replacement for one.

Educational content, not medical advice. Talk to a doctor or pharmacist before starting a supplement, especially if you are pregnant, breastfeeding, or take medication.