Skip to content

Built for researchers

The read you need before you submit.

An accept or reject on your draft, before a human reviewer gives you one.

What it does

Judge

A venue-aware accept or reject on your draft.

Explore

Prior art, your angle, and the gaps still open.

Benchmarked

Scored against the real decisions on 697 papers.

Live sources

Current literature, searched while the review runs.

Venue guidelines

Each verdict applies the venue's own criteria.

Gap Alerts

An email when new papers land in a topic you watch.

How a review runs

Upload a PDF, pick a venue, get back one call against that venue's criteria.

Venue-aware verdict Issues, severities and fixes Results you can reload any time
697
papers benchmarked
0.91
AUC on ICLR 2025
92%
benchmark verdicts matched
17
venues supported

Not only machine learning

Alongside ICLR, NeurIPS and ACL sit eLife, PLOS Medicine, Physical Review Letters and the American Economic Review. Each is read against its own published criteria rather than a single bar we invented: eLife's assessment, for one, rates how important the findings are separately from how strong the evidence is, and the report comes back in those terms.

See it run

These are real runs, not screenshots. Open one and judge the output before you spend anything.

Start a review Back to the overview