Claim Checker

Somebody said it confidently. Is it true?

A curated library of the claims that actually circulate — in forums, in reels, in supplement copy — each with a verdict, the evidence level behind it and the reasoning laid out. Editorial content, written by a human, not generated on the fly.

How to read a verdict. A verdict grades that specific sentence — not the compound as a whole, and not whether it is safe or advisable. The evidence level next to it tells you what kind of studies the verdict rests on. So the same compound can carry a supported claim and a misleading one at the same time — the verdict belongs to the sentence, not to the compound, and never says whether using it is a good idea.

Showing one linked claim.

Supported by evidenceL5Repeated human trialsGLP-1s

“Tirzepatide improves obstructive sleep apnoea in people with obesity.”

Verdict on this exact wording — not on the compound overall, and not on whether use is advisable.

Supported — in the randomised trial populations studied.

Dedicated randomised trials reported reduced apnoea severity.
Effects are tied to the studied populations and treatment durations.
This is a good example of a claim moving from 'promising' to 'supported' because someone ran the trial.

Last reviewed 2026-07-02Compound profile

What the verdicts mean

Supported by evidence: Good human evidence supports the claim as worded, within the populations studied.
Partly supported: There's something real here, but the claim overreaches or blurs a distinction.
Promising but unproven: Plausible mechanism and early data, no controlled human evidence yet.
Not supported as worded: The evidence needed to support this simply doesn't exist. Not the same as disproven.
Misleading: Technically-adjacent to something true, but framed in a way that misleads.

Submit a claim for review →