RO-Crate#

What a RO‑Crate actually is#

  • RO = Research Object – any collection of files that belong together in a research project (data tables, figures, code, PDFs, …).

  • Crate = a container (think of a cardboard box).

A RO-Crate is a standardised digital box that holds all the files plus a small metadata file that tells you what each file is, who created it, when, under which licence, etc. It can be compared to a ZIP folder + a “data sheet” that explains what each element is, who created it, and under what conditions it may be shared.

An attached RO-Crate bundles the actual data together with its metadata, giving you a self‑contained, portable snapshot.

A detached RO-Crate carries only the metadata and points to data that remains stored elsewhere, which keeps the crate lightweight but relies on the external locations staying accessible. Using RO-Crates makes your work FAIR, easier to cite, and ready for journal or archive submission.

The details of how to build RO-Crates are developed openly as a community project.

Quick visual example#

Attached crates keep the data inside the box → fully portable but potentially large.

my-attached-crate/
│
├─ data/
│   └─ measurements.csv          # the file is right here
└─ ro-crate-metadata.json       # "@id": "data/measurements.csv"

Detached crates keep only the labels inside; the data stay outside (usually on a server) → small box, but you rely on the external links staying alive.

my-detached-crate/
│
└─ ro-crate-metadata.json       # "@id": "https://datahub.org/measurements.csv"

Key points of detached RO-Crates:#

  • They exist without defined root directory,

  • They are not processed in a file-system context

  • The referenced data entities are web-based

  • They follow a naming convention: ${prefix}-ro-crate-metadata.json