Check a paper's reproducibility readiness

Give us a bioRxiv or medRxiv DOI — or paste a paper's Methods and availability sections — and we will tell you what it documents about reproducing its analysis: whether versions are pinned, whether the software environment is closed, whether the data and code are reachable. Free, and the result is yours alone.

What this is not.

Nothing is executed. This reads the text you give it and reports a ceiling: an upper bound on what a code check could later confirm. A paper cannot be called reproducible on this basis — only "documented well enough to attempt it". Confirming that the described environment actually holds requires reading the repository, and confirming the results requires running the pipeline.

DOI of the paper — required Required even when you paste the text or upload a file: the report is issued against a paper's identity, and that identity is sealed into a checksum before the assessment begins. A paper with no DOI cannot be assessed here yet. For a bioRxiv or medRxiv DOI this field is all we need — we fetch the article XML from the preprint server. For any other DOI, add the text below.

Paper text — optional for bioRxiv and medRxiv

…or upload a file PDF, Markdown or plain text. A PDF is read with poppler, which recovers reading order across columns — but line-break hyphens do not survive it, so a version like BWA-MEM 0.7.17 can come back mangled. Where the paper is a bioRxiv or medRxiv preprint, the DOI above gives a cleaner read than its own PDF.

I agree that the text of this paper will be sent to Anthropic Claude Sonnet 5 running on Amazon Bedrock in the EU (Frankfurt, eu-central-1) in order to produce this assessment. The text is processed in that region, the assessment is returned to me alone, and the submitted text is discarded once the assessment finishes.

What happens next

For a preprint DOI we fetch the article XML from bioRxiv or medRxiv and read the Methods and availability sections out of it. References, figures and tables are dropped.
The text is assessed against a published rubric on the GenomicsChain anchor. It takes a few minutes.
You get a private link. It is not listed anywhere and we do not publish per-paper results.
Your text is discarded once the assessment finishes; only its checksum is kept.

Curious what this looks like across a whole field first? See the 201-paper bioRxiv survey.