An open library bridging centuries of Islamic knowledge with modern AI — built for researchers, students, and curious minds.
Maktabati.ai vectorises the entire OpenITI corpus — over 8,900 classical works in Arabic, Persian and Turkish — and makes them searchable through modern embedding models.
Each text is chunked, embedded with multilingual-e5-base and stored in a high-performance vector database for RAG applications and semantic research.
Built on an AMD Ryzen 9 5900X · Radeon RX 6900 XT (16 GB VRAM) · 32 GB DDR4 · Ubuntu Server
Fully vectorised OpenITI RELEASE 2025-1 collection of classical Islamic texts, prepared for semantic search and RAG. Each entry represents a text chunk with its embedding vector and complete metadata.