About

Licence facts reported by the dataset and its declared sources. Not legal advice, and not a grant of permission to use the dataset.

Where the facts come from

For each dataset, we check its public Hugging Face metadata and dataset card. We check the names of its top-level files. We read the opening text of matching licence files and look for known licence identifiers. We also check the licence metadata of source datasets listed in the card metadata. Each page shows what these checks recorded.

What the signals mean

• Permissive or share-alike: the checks class the reported licence as permissive or share-alike. They flag no conflict in the metadata or licence text they checked. This does not establish permission to use the dataset.
• Use restrictions: the checks found a licence or card text linked to use restrictions, such as non-commercial or research-only terms.
• Possible licence conflict: the checks flag a difference between the declared licence category and a detected licence file, or between the declared licence category and a source's licence category.
• Unclear: the checks cannot assign another signal. For example, the dataset's licence or a source's licence is missing or not recognised. A missing source is not evidence of permission.

Limits

The signal summarises selected metadata and text on the Hugging Face Hub. The checks do not analyse the full legal terms. They do not verify that a licence declaration is accurate. They do not determine which laws apply. The signal does not establish permission to use the dataset.

Corrections

Found a mistake? Write to hello@datalicense.uk. You can also compare the page with the dataset's current Hugging Face page.