About this site

What this is

Research on ME/CFS, Long Covid and other post-acute infection syndromes (PAIS) is reviewed unevenly. Papers with fatal design problems get published and then get quoted in the press as if settled. Good work sometimes gets ignored. Researchers and patients have limited capacity, ability, and incentive to review each paper deeply.

This site publishes careful AI reviews of individual papers so that anyone can see whether the claims hold up and how strong (or weak) the evidence is.

How a review is made

Each paper is read by Claude against the guide, which is published in full and versioned. Every review records which version it was written under.

About the score

Each review carries a grade out of 10. It is a considered judgement about how much the paper's evidence can carry its claims, not a measurement. Anything below 5 means the paper does not support its main claim.

A low score is not an accusation of bad faith. Most papers that score badly here are ordinary work with design limits that its authors may well acknowledge; the problem is usually how the result then gets used.

What this is not

These are not journal peer reviews, and no journal has commissioned them. They are not a consensus position, and they are not medical advice. They are one careful reading, published openly so it can be argued with. Claude can be confidently wrong, and so can the guide.

If you disagree

Every review has a comment section, and so does the guide. If a review misreads a paper, misses something in a supplement, or is unfair, say so there — that is where it gets argued out first, in the open, where the people who wrote the paper and the people who read it can both see it. Reviews may get corrected; when they do, the edit history is public in the site's repository.