Reproducibility score
One score, backed by concrete findings
A 0–100 score in four categories, so authors, editors and readers see at a glance where a repository stands.
- Poor0–49
- Fair50–69
- Good70–84
- Excellent85–100
Four categories, clear ranges
Reproducibility score
- Excellent85–100Exemplary reproducibility with full automation
- Good70–84Well-documented and structured for reproducibility
- Fair50–69Meets basic reproducibility requirements
- Poor0–49Does not meet minimum reproducibility standards
What sits behind the number
Every finding carries a severity, the evidence and a suggested fix.
- 1Execution
- 2Data & Results
- 3Code Health
- 4Manuscriptwith a manuscript
Execution
Environment, entry points, seeds, traceability and licensing.
Data & Results
Data provenance and integrity, outputs and notebooks.
Code Health
Documentation, clarity, tests, secrets and repository hygiene.
Manuscript
Figures, methods, claims and statistics against the code, when you add a paper.
Works with code in any programming language, including notebooks, environment files, containers and workflows.
Raise your score before you submit
- 01
Start with critical findings
They block others from reproducing key results, so fix them first.
- 02
Apply the suggested fixes
Follow each suggestion, or open drafted fixes as pull requests.
- 03
Analyze again
Re-evaluate with your accepted fixes, or analyze the updated repository.
The same checks run on every analysis, so scores compare well at the same effort level.