Our method

Make the checks visible.

The path from a paper record to a card is automated and documented.

Sources and selection

We search PubMed for exercise, sleep and recovery research. We also use OpenAlex, Crossref and Europe PMC metadata. Source links lead to paper records. The National Library of Medicine does not endorse this site. OpenAlex metadata is CC0; Crossref and Europe PMC records retain their own attribution and terms.

Checks before publication

  1. Every numeric token in our written fields must appear in the source summary.
  2. An independent model assesses whether the claim and mind–body note are supported.
  3. Program rules assign study type from publication types and calculate evidence strength from study design, verified size and certainty wording.
  4. Retraction and correction signals, source links, disallowed wording and the card schema are checked.

A failed card is held back. These checks cannot remove every error or establish personal relevance.

Strength, in plain language

High, moderate, low and very low describe our rule-based reading of the study design and abstract. Meta-analyses with enough verified included studies can rank high; systematic reviews and sufficiently sized randomized trials can rank moderate; observational work usually ranks low. Low-certainty wording lowers the rating. A strength badge is not a promise.

Corrections

If you find an error or a changed paper status, email [email protected] with the card ID and paper link. We will remove or correct a card and record its version.