RASD, the Research Unit in Arabic and Social Digital Spaces, is part of the Arab Center for Research and Policy Studies (ACRPS). The unit advances Arabic digital capabilities through open-source natural language processing, machine learning, and computational linguistics tools, with a focus on language and regional contexts. Its projects include DALLA, a collection of Arabic large language models; NeoAraBERT, an Arabic embedding model; and ArabicMMLU and other benchmarking work for Arabic NLP, cultural awareness, and local value systems.

RASD also develops knowledge graphs, curated multimodal datasets, and tools for research and publishing. These include Arabica Wikidata resources, an editorial tool suite that automates tasks such as citation parsing and index creation, and a visualization framework for the Arab Opinion Index. Additional projects include Hawir Kitabak, an AI-powered author chat experience on Mastodon, and a custom URL shortener for ACRPS.

RASD shares research publications, datasets, code, and interactive resources, including Colab tools for Arabic relation extraction and topic classification. Its work supports researchers and organizations developing Arabic-language technologies and digital content