Paste a report with its references. Every citation gets looked up in the literature, and every claim gets read against the paper it names, including the tables and the figures.
Check a reportMitchell and Popham (2008), analyzing mortality records for over 40 million English residents, found that income-related health inequalities were narrower in greener areas.
Says 40 million mortality records; paper analysed 366,348
Li (2010) reported that weekend forest walks increased natural killer cell activity by 56% and that the effect persisted for 30 days.
No 56% figure in text; exposure was 3-day/2-night trips
The subjects experienced a 3-day/2-night trip to forest areas… the increased NK activity lasted for more than 30 days after the trip.
▸ 31 with nothing to flag
Every one of these is real, from a report we checked. None of them look wrong on the page.
Kondo et al. 2018
“Kondo et al. (2018) conducted a cluster-randomized trial in Philadelphia in which vacant lots were randomly assigned…”
Cited paper is a systematic review, not a randomized trial
Mitchell & Popham 2008
“…argued that green space acts as an ‘equigenic’ environment, one that reduces health inequality.”
Quoted term “equigenic” does not appear in this paper
Zhu et al. 2013
“An uncommon variant of CES1 encodes the nonsynonymous substitution Gly143Gln…”
Paper reports Gly143Glu, not Gly143Gln
01
Against CrossRef, PubMed and OpenAlex. Wrong DOIs get corrected, and papers that don’t exist get named as such.
02
Open access where we can reach it, and you can upload the rest. You see which sources are readable before anything is checked.
03
Prose, tables and figures. Anything that doesn’t match comes back with the sentence, the reason, and the passage it was checked against.
It doesn’t rewrite your work. Every finding is labelled and left for you to decide on. Nothing is edited, and nothing is applied on your behalf.
It can’t read what it can’t reach. Paywalled papers are checked against their abstracts unless you upload a copy, and the report says which ones those were.
It doesn’t judge whether you’re right. It checks whether the paper you named says what you said it does. That’s a narrower question, and the only one it can answer honestly.