Report · Live

Bayanat Labs Open-Programme Policy and Operating Manual

The rules Bayanat Labs follows when it chooses, licenses and publishes open Arabic data, as files anyone can read, check and reuse: the open-programme policy (selection rubric, release tiers, licence matrix), the operating manual with its gated stages and definition of done, and bl-score, a command-line tool that applies the rubric to a scoping memo. In English and Arabic. Documentation, not legal advice.

By Bayanat Labs· Updated · CC BY 4.0· اقرأ بالعربية
LicenceCC BY 4.0 (text); Apache-2.0 (code)
Versionv1.0.1 · Policy v1.0.1 and operating manual v1.0.1; bl-score v0.1.0.
Published
Updated
Who it is forTeams that publish open Arabic data and want a licence-first release policy and a gated build process to adopt or adapt.
LevelIntermediate

open data policy Arabic dataset licensing dataset release process dataset selection rubric data provenance

Key numbers

13stages in the operating manual
12items in the definition of done

Counted by script from the released files. The repository has no MANIFEST.json. Each key number is counted from the operating manual as published on /open/how-we-build and pinned to its SHA-256.

Downloads and identifiers

Get started

Run the policy checks and bl-score locally

git clone https://github.com/Bayanat-Labs/open-policies
cd open-policies
python -m venv .venv && . .venv/bin/activate
pip install pyyaml jsonschema pytest
pip install ./tools/bl-score
python policy/scripts/validate_policy.py --policy-dir policy
bl-score tools/bl-score/examples/green_p0.yaml

What is released

PartLicenceIn this release
Open-programme policy: text, selection rubric, release tiers, licence matrix and schemas (policy/)CC BY 4.0Yes
Operating manual: public edition in English and Arabic, stages, definition of done and crawler rules (manual/)CC BY 4.0Yes
bl-score, validators, scripts and testsApache-2.0Yes

Attribution: Policy by Bayanat Labs; Operating manual by Bayanat Labs. https://github.com/Bayanat-Labs/open-policies

Intended uses

  • Read the rules behind every Bayanat Labs open release: release tiers, licence matrix, selection rubric and gated stages.
  • Adopt or adapt the licence matrix and the operating manual for your own open-data programme, with attribution.
  • Score a dataset idea against the rubric with bl-score.

Out of scope

  • Legal advice for your organisation: the files are documentation, not a legal opinion.
  • A substitute for checking the licence and terms of each source you use.

Rights at a glance

Training, display, redistribution and synthetic-voice use are separate permissions.

UseStatusNote
TrainingAllowedText under CC BY 4.0, code under Apache-2.0.
DisplayAllowedCC BY 4.0, with attribution.
RedistributionAllowedText under CC BY 4.0 with attribution and a note of any changes; code under Apache-2.0.
Synthetic voiceNot applicableText and code only.

Limitations

  • Written for Bayanat Labs' own releases: the thresholds and gates are its defaults, and a project may make them stricter.
  • Statutes and platform terms are summarised only to explain the rules; check the current text of each before relying on it.

Sources and licences

Written by Bayanat Labs. The files cite statutes and platform terms by link and short quotation only, to explain the rules.

How to cite

Cite as:

Bayanat Labs (2026). Bayanat Labs open policies, version 1.0.1. Bayanat Labs. https://github.com/Bayanat-Labs/open-policies

No DOI. Cite the URL in the citation above and the date you read it.

@misc{bayanatlabs2026openpolicies,
  author       = {{Bayanat Labs}},
  title        = {Bayanat Labs open policies, version 1.0.1},
  year         = {2026},
  month        = oct,
  howpublished = {\url{https://github.com/Bayanat-Labs/open-policies}}
}

How we handle citations

Changelog

DateVersionChange
v1.0.1Published on GitHub: policy v1.0.1, operating manual v1.0.1 and bl-score v0.1.0, in English and Arabic.

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

Need production-grade, rights-cleared data?

We publish synthetic fixtures, open benchmarks, results, code and guides for free. Data at scale is licensed or built to order: ready-to-license Arabic speech and text datasets with commercial AI-use rights written into the agreement, and custom collection, annotation and evaluation.

Bayanat Knowledge: License books, subject-organized MSA text, and structured question-answer data. Each collection has a clear training task, with its own counts and licensing details.