New partnership

Licensed Arabic Academic Text Corpus

Rights-cleared Modern Standard Arabic (MSA) academic texts for AI and LLM training, delivered via Dialect Data.

Overview

This corpus is licensed in partnership with Al-Maktab Al-Arabi Lil-Maaref (Cairo). It is designed for teams building Arabic LLMs who need high-quality text with clear commercial usage rights and licensed access via Dialect Data.

Pilot highlights

~70 academic titles
~4.3M words
Domains: psychology, social sciences, law, economics, technology
Commercial AI/LLM usage rights included
Non-exclusive license

Use cases

Pretraining and continued pretraining
Fine-tuning and instruction tuning
Arabic evaluation / benchmarking datasets
Domain adaptation for academic and formal Arabic

Delivery format

Delivered as one file per book (TXT, UTF-8), organized by section, plus a structured metadata sheet. Title-level details are shared under NDA.

Exact file naming convention and delivery structure are standardized for ingestion pipelines.

Rights and compliance

Rights-cleared for commercial AI use
No stand-alone redistribution of texts
Access provided via controlled delivery and approvals

Request access

Tell us your use case (training, eval, domains, size). We can share access details and title-level details under NDA.