APPLIED AI GOVERNANCE · QUALITY ENGINEERING
Documentation that
stands up to scrutiny.
Audit and compare public Hugging Face Model Cards, Dataset Cards and Spaces. Reveal missing evidence signals, prioritize improvements, and export transparent reports, directly from your browser.
WORKBENCH / 01
Run a documentation audit
Choose public Hub repositories or paste a card for private, offline analysis.
One per line · up to 5
Use prefixes
model:, dataset:, or space:. Full Hugging Face URLs work too. The default for unprefixed IDs is Model.Ready. Choose an audit mode to begin.
AUDIT RESULTS / 02
Evidence dashboard
Heuristic documentation coverage across selected repositories.
REPOSITORIES0audited records
AVERAGE COVERAGE0%documentation presence
DETECTED SIGNALS0/0rubric checks passed
OPEN PRIORITIES0missing signals
Repository comparison
CLICK TO INSPECTScores measure documentation signals only. Differences can reflect card types with different criteria.
Documentation details
—
Remediation studio
Prioritized actions for the selected card, with safe placeholders.
METHOD / 03
A transparent scoring rubric
Each card type has a published, 100-point checklist. A signal is detected if the document contains the matching heading or structured YAML metadata. A detected heading does not mean its content is complete or accurate.
03artifact types
100points per rubric
0external AI calls