Creatine used to live in the gym bag. Now it’s sold for your brain too.
A survey of 2,178 creatine users published this year found that two thirds take it for mental performance. About one in three of them noticed it working.
So who’s right, the label or the users? The research exists, a few hundred papers of it, and reading it is the slow part. We asked an agent to read the abstracts and check which studies get cited most.
Two calls, under a cent
The agent searched two research databases through Lightning Enable’s Agent Commerce Store, which sells agents per-call access to them:
- PubMed, for creatine and cognition: 194 matches, and the newest 49 came back with their abstracts split into methods, results and conclusions. $0.0046.
- OpenAlex, for the same topic: 50 papers with their citation counts and abstracts, and a link to free full text wherever one exists. $0.0023.
Together, $0.0069. Then the agent read the abstracts and compared the citation counts. It didn’t open the full papers.
On a rested brain, little changes
In a trial published this year, 40 young, active adults took 20 grams a day for a week. Their average power in a 30-second cycling sprint rose 8%. Their scores on every cognitive test stayed where they were.
A 2023 trial gave young adults 10 or 20 grams a day for six weeks, with the same result for cognition. A 2008 study put it in its title: “Creatine supplementation does not improve cognitive function in young adults.”
On a tired brain, it shows up
The effect appears when the brain runs short of energy.
In a study published in April, 29 people stayed awake for 21 hours. Those who took one dose of creatine (0.2 grams per kilo of body weight) kept more of their performance on logic, number and reaction tests than those on placebo, by up to 12%.
In 2015, 15 adults breathed air with 10% oxygen for 90 minutes. After a week of creatine, their brain creatine was up 9.2%, and the attention they lost to low oxygen came back.
A 2026 review of the sleep-deprivation research found five studies with a positive trend, and asked for larger trials.
Vegetarians, and older adults
Most dietary creatine comes from meat and fish. In 2010, 128 young women took 20 grams a day for five days. Memory improved in the vegetarians among them.
The widest view comes from a 2024 meta-analysis of 16 randomized trials with 492 people aged 21 to 76. Memory improved by a small margin (an effect size of 0.31), and attention and processing speed improved too. Overall cognition and executive function stayed the same.
For healthy older adults, a 2026 review calls the evidence limited and points towards a benefit. For dementia, a 2026 scoping review screened 8,317 records and found one human pilot trial, with 20 patients. The randomized trials are still to come.
What the agent concluded
Creatine helps the brain most when it’s short of energy: no sleep, thin air, a diet without meat. For a rested brain on a mixed diet, the trials mostly find no change, which matches what most of the surveyed users reported.
The researchers agree on the gap. In October, a review in Experimental Physiology asked in its title why creatine’s “physiological plausibility” has yet to turn into convincing clinical evidence.
This is a summary of published research. For your own health, ask your doctor.
Read the studies yourself
Of the 14 papers in the OpenAlex results cited at least 50 times, 8 have free full text. The agent linked them:
| Study | Year | Cited by | Free full text |
|---|---|---|---|
| Creatine metabolism and psychiatric disorders (Allen) | 2012 | 186 | Europe PMC |
| Creatine supplementation and brain health (Roschel et al.) | 2021 | 174 | Nutrients |
| Creatine and cognition in healthy people, a systematic review (Avgerinos et al.) | 2018 | 165 | PubMed Central |
| Creatine and cognition during oxygen deprivation (Turner et al.) | 2015 | 157 | J Neurosci |
| Creatine and brain function and health (Forbes et al.) | 2022 | 154 | Nutrients |
| Creatine, strength training and cognition in older women (Alves et al.) | 2013 | 109 | PLOS ONE |
| Creatine in vegetarians and omnivores (Benton, Donohoe) | 2010 | 101 | Cambridge |
| Creatine and cognition in adults, a meta-analysis (Xu et al.) | 2024 | 53 | Frontiers |
Two of these are paywalled at their journal. The agent found the free copy on PubMed Central and Europe PMC. For the six without free full text on record, including that 2008 study, it links the publisher’s page.
Use TaskFuel to search PubMed and OpenAlex for studies on magnesium and sleep. Tell me what the randomized trials found and who benefited, and link free full text for each key study where it exists.
What else your agent can look up
Lightning Enable’s Agent Commerce Store offers more than search. Also in the TaskFuel catalog, with prices quoted before purchase:
- A researcher’s profile on OpenAlex, with their works, citations and h-index: $0.0023
- Detailed paper metadata from OpenAlex, with topics and links to available full text: $0.0038
- One PubMed article with its structured abstract and MeSH terms: $0.0054
- US Census population for a place name like “Austin, TX”: $0.0023
- US Census demographics for the same kind of place: $0.0038
Good to know
A search returns up to 50 results in one call, and PubMed returns the most recently added first. To reach older studies, narrow the search, for example to meta-analyses. OpenAlex had no abstract for 6 of the 14 most-cited papers. For those, the agent had the title, the citation count and the link.
Your agent reads the research
With TaskFuel, your agent searches the literature, reads the abstracts and checks who cited what, in the same conversation where you asked the question. It can go from PubMed to a researcher’s citation record to Census data as your questions change, with one TaskFuel account behind all of it.
Connect your agent and the first $1 is on us.
Check the research on [your supplement] for [what you take it for]. Search PubMed and OpenAlex through TaskFuel, tell me who benefited in the trials, and link free full text where it exists.