Is This Claim True?
Paste the claim. Get a verdict, the corrected version and the reason in about five seconds.
A statistic, a quote, a cited study: one model checks it against what it knows and tells you whether it is safe to publish. Free, once a day. Paid plans put four models on the same claim and show you where they split.
This usually takes 3-5 seconds.
Deep takes a minute or two: four pro models answer, two critics review, one mediator writes the final call. Leave this page open and the result lands here.
Claim Checker runs the moment you sign in, on the text you just pasted. Nothing to enter twice.
A saved check is kept for 24 hours.
You've used up today's free runs on this check.
Come back tomorrow for another free run.
Why Check Claims Before Publishing?
AI-generated content has a 17-34% claim error rate. The claims that look most confident are often the ones most likely to be wrong — AI uses 34% more confident language when hallucinating.
Paste your claim
Any statistic, quote, factual assertion, or sourced claim you want to verify.
AI evaluates the evidence
The model checks the claim against known facts, assesses whether cited sources are real, and flags overclaiming.
Get a clear verdict
Verified, Partially Verified, or Challenged — with the correct version and an explanation of what's wrong.
Research shows 47% of ChatGPT-generated citations are inaccurate or completely fabricated (PMC 2024). Checking claims before publishing prevents the correction email nobody wants to send.
This free tool checks one claim at a time. To fact-check a whole draft across four models at once, see the TrueStandard AI fact checker. Checking a source, not a number? The AI citation checker resolves the DOI and reads the page.
Example Check
"Intermittent fasting reduces inflammation by 70%"
The trials behind this measured CRP, one marker, and found a modest drop in it, not in 'inflammation' broadly. The 70% figure appears nowhere in the peer-reviewed literature. Health content overclaims this way often.
Clinical trials of intermittent fasting found a modest drop in CRP, one inflammation marker.
No specific source is cited. The claim appears to conflate multiple studies with different findings.
Frequently Asked Questions
Paste the claim: a statistic, a quote or a number. One model checks it against what it knows and returns a verdict: verified, needs fixing or challenged. If it needs fixing, you get the corrected version and the reason.
ChatGPT tends to agree with claims you present to it — a phenomenon called sycophancy. This tool is specifically prompted to challenge claims, not confirm them. It checks evidence, not vibes.
A single model catches obvious errors but has blind spots. The free run uses 1 model. Council puts 4 independent models on the same claim and caps the confidence by how much they agreed. Review does that for every claim and citation in a draft.
We store your claim and the verdict so your receipt link keeps working; the receipt is private unless you share it. We use enterprise API agreements, so your content is never used for model training.
This tool checks 1 claim with 1 model. Review checks every claim and citation in your draft with 4 models, shows where they disagree, and gives you a shareable report. It's the difference between checking one ingredient and having the whole dish reviewed.
Powered by TrueStandard
Four independent models check every claim in your draft.