Supplements work [1–5]. Add in the prerequisites: for certain people, for certain outcomes, at certain doses. The baseline holds. A lot of supplements work. Not just in theory: there is human research with positive effects on outcomes a person would reasonably care about—from attention in older adults to respiratory symptoms in athletes.

Finding out what an individual supplement supports and when it helps is straightforward, whether you turn to evidence based podcasts or articles, research papers, or AI prompts. You search for what would best support you and your goals, use a few sources, find a few supplements.

You found some. Supplements you never found don’t leave empty spaces in your hand. A different podcast, a different paper, a different rabbit hole, and you're holding a different mix of pills, with a perfectly good explanation for taking those instead.

How much is there to miss? The research extends across more than 1,500 supplement interventions. Take one, arachidonic acid, a fatty acid you’ve probably never heard of. One 2019 review synthesized findings from fourteen trials that ranged from muscle function to blood clotting [6]. Not fourteen studies in animals or petri dishes—there are far more of those—fourteen randomized controlled trials in humans.

Suppose the sources you trust have compared the relevant research and identified the best supported supplements for your goals. One has better results, but the evidence is concentrated in people less like you. Another has more modest results in circumstances much closer to your own. You weigh that up and settle on one or both, then turn to the available products. You find some, and choose a bottle from a reputable brand, or at least one you like.

Maybe you settled on a lower dose for a simpler formula, or a few ingredients you weren’t looking for to get the dose you wanted; you found something good enough, and now you can start taking it.

The whole formula gets filed under what you bought it for. You can give a perfectly reasonable answer without remembering everything you weighed and every compromise you made along the way. You carry forward your reasons for taking each bottle.

Your supplements probably work. I have good reasons for taking mine, and the research to back them up. A few years ago, I tried to use that same research to answer a slightly different question: given what I already take and what I actually want to improve, what is most worth adding?

Works and worth have three letters in common. I was optimistic about the last two.

I started with a chatbot that could retrieve research and use it alongside a person’s goals and existing supplements. I figured that more primary research, more personal context, more products to search from, and better analysis at each of those steps would close that gap. I kept improving the search, extracting the findings to give it more useful detail from more papers, and adding tools to check products and interactions. The answers genuinely improved.

The AI now had significantly more of the relevant research to compare, but still only some of it. Then what was found guided the product search, which returned some of the available formulations. Whatever the inputs, it could only compare what it found. A better supported supplement or a better suited product could still be missing from the decision.

The answer drifted further when considering everything the person already took. I needed to know what the new product added to a routine already full of things with evidence behind them. That required the same evaluation for every ingredient already there, then an account of what they supported together and how strongly. It could explain why each bottle might help, just as you could. But, the accounts didn't carry enough depth to be the basis for deciding what was worth adding. Give one ingredient too much credit and a worthwhile addition looks redundant; miss its contribution and more of the same looks like progress.

I came to realize these bottlenecks weren't unique. Suppose the sources you trust have compared the relevant research and identified the best supported supplements for your goals.The supplements we take aren't words from a research paper. They're physical capsules, softgels, or tablets from a physical bottle we buy on top of more we're already taking. I didn't want to leave my question unanswered, so I built the solution from the ground up.

Starting with the evidence. More than 8,000 human trials and reviews, with findings tied directly to specific outcomes, populations, doses, and comparators. That evidence is weighed across every supplement relevant to a person’s priorities, focused on outcomes from populations similar to theirs methodology. This means no viable option is overlooked from the start, and every comparison is based on results found in individuals like them. That is directly mapped to the ingredients in more than one hundred thousand products. The evidence is applied to the whole formula—its forms, amounts and other active ingredients at the proposed serving—so what the research supports for you is based on what you can actually take product selection.

To answer what was most worth adding, I put everything already being taken through that same analysis. For each person, the analysis establishes what their existing products support, how strongly, and where that support overlaps. Each feasible addition is evaluated for what it would contribute alongside them.

Those last two letters. Took the last three years.

Today, you can use that same comparison to understand what is worth keeping or changing. The score reflects how your stack compares with the evidence supported opportunities available for your goals and circumstances. Its breakdown shows which parts of your goal are supported and what each product contributes alongside the others.

You decide what matters to you, what you’re willing to spend, and which trade offs you want to make. You should be able to make those choices with the comparison in front of you, rather than first having to assemble it from separate partial interpretations of the research, the products and everything else you take.

I’d started by asking what was most worth adding. Answering that meant building a way to make every paper and endpoint as useful and actionable as possible. Now you can answer: what is worth taking.

Supplement AI available on the web and in the App Store. Start with the supplements you already take and what you actually want to improve. A few minutes of setup brings that analysis to your own stack.

We don’t sell supplements, run ads or take affiliate commissions. Buying more supplements doesn’t pay us more.

Your first stack analysis and optimizations are free.

Selected examples, chosen to illustrate the breadth of positive findings across supplement interventions, populations, and outcomes.

Rodrigo-Gonzalo, María José, et al. “Effect of Polyphenolic Complements on Cognitive Function in the Elderly: A Systematic Review.” Antioxidants, vol. 11, no. 8, 2022, article 1549. doi:10.3390/antiox11081549.
Yao, Jia, et al. “Efficacy of Flavonoids-Containing Supplements on Insulin Resistance and Associated Metabolic Risk Factors in Overweight and Obese Subjects: A Systematic Review and Meta-Analysis of 25 Randomized Controlled Trials.” Frontiers in Endocrinology, vol. 13, 2022, article 917692. doi:10.3389/fendo.2022.917692.3
Möller, Gabriella Berwig, et al. “Supplementation of Probiotics and Its Effects on Physically Active Individuals and Athletes: Systematic Review.” International Journal of Sport Nutrition and Exercise Metabolism, vol. 29, no. 5, 2019, pp. 481–492. doi:10.1123/ijsnem.2018-0227.
Akuffo, Kwadwo Owusu, et al. “The Impact of Supplemental Antioxidants on Visual Function in Nonadvanced Age-Related Macular Degeneration: A Head-to-Head Randomized Clinical Trial.” Investigative Ophthalmology & Visual Science, vol. 58, no. 12, 2017, pp. 5347–5360. doi:10.1167/iovs.16-21192.
Dang, Chun, et al. “Chinese Herbal Medicines for the Treatment of Depression: A Systematic Review and Network Meta-Analysis.” Frontiers in Pharmacology, vol. 15, 2024, article 1295564. doi:10.3389/fphar.2024.1295564.
Calder, Philip C., et al. “A Systematic Review of the Effects of Increasing Arachidonic Acid Intake on PUFA Status, Metabolism and Health-Related Outcomes in Humans.” British Journal of Nutrition, vol. 121, no. 11, 2019, pp. 1201–1214. doi:10.1017/S0007114519000692.