Bayanat Labs Arabic Dataset Rights Index
Can a commercial model be trained on this Arabic dataset? Can copies be shared? Does the licence ask for share-alike? For 120 datasets on Hugging Face (the 100 most-downloaded datasets tagged Arabic, and 20 more repositories in sets of known copies), the index answers from the declared licence with the rules of the Bayanat Labs licence checker and from what the dataset card says about use restrictions, and adds what the card says about extra terms, upstream sources, personal data and access. When a card limits use to research, non-commercial or personal use, or says the underlying data keep their own licences, an answer that would permit is unclear. The index also links repositories that hold copies of the same data and flags copies that declare different licences. Each answer quotes its evidence with the line, the commit read and the retrieval date. Assessed from the declared licence and card text on 19 October 2026; not legal advice; check the source terms.
| Licence | CC BY 4.0 (verdicts, flags, notes and documentation); CC0 1.0 (Hub metadata facts); Apache-2.0 (code); quoted card text stays under its owners' terms |
|---|---|
| Version | v0.1.0 · Index and code v0.1.0, assessed on 19 October 2026; the code is tagged on GitHub and not on PyPI yet. |
| Published | |
| Updated | |
| Who it is for | Teams that pick Arabic training or evaluation data from Hugging Face and need to see, before they use a dataset, what its declared licence says about commercial training and redistribution, what its card says about extra terms and personal data, and whether its copies agree. |
| Level | Beginner |
Key numbers
Counted by script from the released files. Each key number is read from stats.json in the Hugging Face dataset and pinned to its SHA-256; MANIFEST.json lists every file with its SHA-256.
Downloads and identifiers
| Where | What |
|---|---|
| Hugging Face | Search the index |
| bayanatlabs.com | Can I use this Arabic dataset? |
| Hugging Face | Dataset on Hugging Face |
| huggingface.co | Verdicts table (CSV) |
| huggingface.co | Evidence table (CSV) |
| GitHub | Code and method on GitHub |
| GitHub | How the answers are computed |
Get started
Load the tables with the datasets library, or install the command line from GitHub at the v0.1.0 tag to look a dataset up; to rebuild the index, install from a clone of the repository, which holds the search configuration. The package is not on PyPI yet.
# Python
from datasets import load_dataset
verdicts = load_dataset("Bayanat-Labs/arabic-dataset-rights-index", "verdicts", split="train")
evidence = load_dataset("Bayanat-Labs/arabic-dataset-rights-index", "evidence", split="train")
# Look a dataset up
pip install "git+https://github.com/Bayanat-Labs/arabic-dataset-rights-index@v0.1.0"
hf download Bayanat-Labs/arabic-dataset-rights-index --repo-type dataset --local-dir rights-index
bl-rights lookup rights-index pain/MASC
# Rebuild the index from a clone
git clone --branch v0.1.0 https://github.com/Bayanat-Labs/arabic-dataset-rights-index && cd arabic-dataset-rights-index && pip install .
bl-rights harvest --config config/search.yaml --raw raw/
bl-rights build --config config/search.yaml --raw raw/ --out build/What is released
| Part | Licence | In this release |
|---|---|---|
| Verdicts, flags, notes and copy groups (Parquet and CSV) | CC BY 4.0 | Yes |
| Hub metadata facts: repository ids, URLs, commits, counts, declared licence tags, file names and dates | CC0 1.0 | Yes |
| Short quotes from dataset cards and licence files | Each quote stays under its owner's terms and links to its line at the commit read | Yes |
| Builder and command line (harvest, build, lookup), for Python 3.10 or later | Apache-2.0 | Yes |
| Search page on Hugging Face Spaces, in English and Arabic | Apache-2.0 | Yes |
| IBM Plex Sans Arabic font in the search page, unmodified (SIL Open Font License 1.1) | SIL Open Font License, copyright IBM Corp. | Yes |
Attribution: Arabic Dataset Rights Index by Bayanat Labs, CC BY 4.0 (verdicts, flags and notes). https://huggingface.co/datasets/Bayanat-Labs/arabic-dataset-rights-index
Intended uses
- Check, before you use an Arabic dataset from Hugging Face, what its declared licence says about commercial training, redistribution and share-alike, with the quoted evidence.
- Find the copies of a dataset on the Hub and see whether they declare the same licence.
- Count licences and card flags across popular Arabic datasets, citing the version and the assessment date.
Out of scope
- Legal clearance: the answers are information, not legal advice.
- Settling the rights in the data itself: the index reads the declared licence and the card, not the sources a dataset was built from.
Rights at a glance
Training, display, redistribution and synthetic-voice use are separate permissions.
| Use | Status | Note |
|---|---|---|
| Training | Allowed | Verdicts, flags and notes under CC BY 4.0 with attribution; Hub metadata facts under CC0 1.0. Quoted card text keeps its owners' terms. |
| Display | Allowed | CC BY 4.0 with attribution; keep each quote with its source link. |
| Redistribution | Allowed | Verdicts and notes under CC BY 4.0, metadata facts under CC0 1.0, code under Apache-2.0, the bundled font under the SIL Open Font License; quotes keep their owners' terms. |
| Synthetic voice | Not applicable | Metadata only; no audio. |
Limitations
- The answers concern the declared licence and the card; a dataset built from sources with their own terms can carry a licence its sources do not allow.
- Verdicts are computed by fixed rules and are not individually reviewed; a wrong or missing licence tag on the Hub gives a wrong answer here.
- A declared licence outside the rules is reported as not covered, and the answer is unclear. Gated datasets are not opened.
- Card statements, including use restrictions, are found by fixed patterns, so some are missed and some quotes may not be about the dataset; each quote links to its line so it can be checked.
- Copies are linked by identical files, card references, a shared homepage or paper; copies with changed files and no shared link are not found.
- Download counts and licences change; each row records the commit and time it was read. Install the code from GitHub for now; the package is not on PyPI yet.
Sources and licences
| Source | Licence | Redistributed here |
|---|---|---|
| Hugging Face Hub: public dataset metadata, dataset cards and licence files | Metadata facts recorded under CC0 1.0; card and licence text quoted in short spans, under its owners' terms | Yes |
| Dataset Licence Compatibility Checker (license-check) 0.1.0, rules version 2026.10.19 | Apache-2.0 (code); CC BY 4.0 (rules) | No |
Each verdict is computed from the licence declared on the Hub with the rules of the licence checker, and is unclear when the card restricts use or says the underlying data keep their own licences; each flag points to a quoted line of the card or licence file at the commit read, with the retrieval time. Requests to the Hub were spaced at least one second apart and carried the Bayanat Labs open-data user agent; gated datasets were not opened and no dataset files were fetched. Source details
How to cite
Cite as:
Bayanat Labs (2026). Arabic Dataset Rights Index, version 0.1.0, assessed 2026-10-19. Bayanat Labs. https://huggingface.co/datasets/Bayanat-Labs/arabic-dataset-rights-index
No DOI. Cite the URL in the citation above and the date you read it.
@misc{bayanatlabs2026arabicdatasetrights,
author = {{Bayanat Labs}},
title = {Arabic Dataset Rights Index, version 0.1.0, assessed 2026-10-19},
year = {2026},
month = oct,
howpublished = {\url{https://huggingface.co/datasets/Bayanat-Labs/arabic-dataset-rights-index}}
}
Changelog
| Date | Version | Change |
|---|---|---|
| v0.1.0 | Index of 120 datasets on Hugging Face (the top 100 tagged Arabic and known copies), assessed on 19 October 2026; search page on Hugging Face Spaces; code and method on GitHub. |
Corrections and removal requests
If something in this release should not be there, or is wrong, email hello@bayanatlabs.com. We handle removal requests within 30 days, and platform requests within 24 hours. How removal requests work
Related product: Bayanat Knowledge