Turn any paper into a Croissant file

Upload an academic paper that introduces an ML dataset. An AI agent reads the PDF, extracts every piece of dataset metadata it can find, and produces a validated MLCommons Croissant JSON-LD file — the emerging standard for machine-readable dataset metadata.

Upload PDF

Drop an academic paper describing a dataset

AI reads it

Agent extracts names, splits, fields, license, and more

Validates

Checks against the Croissant schema with mlcroissant

Download JSON-LD

Get a ready-to-use Croissant metadata file

Drop a PDF here or click to browse

Academic paper describing an ML dataset (max 15 MB)

You'll be notified when the job completes.