Arabic AI Atlas

•
45
•

Navigate, query, and recommend open models, datasets, benchmarks, and tools across the Arabic artificial intelligence ecosystem.

Arabic AI Atlas is a curated, comprehensive map of the Arabic AI ecosystem that can be integrated directly into Claude Code and other MCP-compatible agents. Drawing from thousands of categorized entries across Arab countries and international contributors, it indexes Arabic large language models, speech recognition (ASR), text-to-speech (TTS), OCR, embeddings, datasets, and benchmarks. Through its embedded offline MCP server and agent skills, it empowers AI assistants to deliver tailored recommendations, track model lineages, and identify ecosystem gaps based on dialects, deployment constraints, and licensing needs.

Características Principales

01Built-in Model Context Protocol (MCP) server providing offline queries and model lineage tracking
02Comprehensive directory covering thousands of Arabic LLMs, ASR, TTS, OCR, datasets, and benchmarks
0343 GitHub stars
04Direct integration with Claude Code via installable marketplace skills and plugins
05Automated tracking of Hugging Face downloads, community updates, and missing ecosystem gaps
06Recommendation engine to match requirements like dialects, mobile runtime, and open licenses

Casos de Uso

01Equipping coding agents with an offline catalog to recommend compatible datasets and Arabic NLP tools during development
02Discovering specialized Arabic speech-to-text or TTS models optimized for specific regional dialects or edge devices
03Tracing model family lineages and base architectures across open Arabic LLMs and embeddings