Heard a health claim?
Check it against the research.

Paste it in. We read the studies on PubMed and weigh how strong they are.

Not medical advice. Ask a doctor before you change a treatment.

Enter one claim. Add the group or situation if it matters.

Previously checked claims appear as you type. Use the arrow keys to choose one, and Enter to open its saved answer.

Don't include names or personal medical details. Claim text is stored and sent to our AI provider. Privacy details

Or try one of these:

Studies read
90
Claims checked
18
Pooled analyses
38

Latest checks

  1. No. 20 Apple cider vinegar cures diabetes Likely false
  2. No. 19 Creatine improves strength Likely true
  3. No. 18 exercise helps depression Likely true
  4. No. 17 creatine builds strength Likely true
  5. No. 16 Vitamin C prevents colds It's complicated
  6. No. 15 apple cider vinegar helps diabetes It's complicated
  7. No. 14 Apple cider vinegar causes diabetes Not enough evidence
  8. No. 13 Eating carrots helps you see in the dark It's complicated
  9. No. 11 Reading in dim light damages your eyes Not enough evidence
  10. No. 10 Eating eggs raises your risk of heart disease It's complicated

Before you trust a verdict

Where do the studies come from?

From PubMed, the U.S. National Library of Medicine's index of medical research, with more than 37 million records. A new check searches it and reads the abstracts of up to 8 of the closest matches. Every study it read is listed under the verdict, with a link to its PubMed page.

Is an AI making this up?

An AI model (Google's Gemini API) reads those abstracts and writes the verdict. It can only cite studies it was shown: the code checks every citation against that list. A verdict with no citations, or one resting only on case reports or animal and lab studies, is turned into “Not enough evidence” before you see it. It can still misread an abstract, which is why every study it used is one tap away.

Is it biased? Who made it?

Evident is an independent project. It shows no ads, sells nothing and has no accounts, so no verdict makes it money. Its rules lean towards caution: “It's complicated” and “Not enough evidence” are ordinary answers here, and a confident one needs studies behind it. Research has blind spots of its own, so every verdict also says what is still open.

Is it free?

Yes. No account, no paywall and no ads. It runs on the free tier of an AI service, so on a busy day it can reach its limit, and the page then says so and when to try again. A claim someone has already checked opens straight away. On that free tier, Google's terms let it use the text you paste to improve its products, so leave out anything personal.

Is this medical advice?

No. It tells you what published research says about a claim, which is not the same as what is right for you. Take what you find to a doctor or pharmacist who knows your history.

Can I see a real one?

“Apple cider vinegar cures diabetes”

It's complicated

Apple cider vinegar can lower blood sugar, but nothing shows it cures diabetes.

Checked 22 September 2026. 8 studies read.

Looking it up on PubMed.

Their contact details are usually on the PubMed page. Most researchers don't answer cold emails, but some do.

How a check works

  1. Your claim becomes a search. An AI model turns "apple cider vinegar cures diabetes" into the terms researchers actually use.
  2. PubMed answers. That's the U.S. National Library of Medicine's index of medical research, 37 million papers. We take up to 8 of the best matches with their abstracts.
  3. The studies get weighed, not counted. A meta-analysis of many trials outranks one small pilot. A case report about one patient barely counts. The model reads the abstracts and says which studies it relied on.
  4. The verdict is stamped, with rules. "False" needs studies that contradict the claim. Nothing that tests the claim, or only single case reports and lab work, means "not enough evidence," never a confident answer. "It's complicated" is for real evidence that is mixed or uncertain. Studies that only match the words don't count.
  5. Every verdict says what's still open. Research is rarely finished. The last line names the gap: who the studies left out, how short they ran, or what finding would change the answer.

The highlighted line says what that evidence shows about the claim itself. Strong evidence of a small effect is not strong evidence for a cure.

Not medical advice. The model can misread an abstract, and PubMed doesn't index everything. Read the studies. Talk to a clinician about your own health.