Convert Arabic PDF books into searchable, accessible, AI-enhanced digital libraries. Most OCR tools mangle Arabic's right-to-left script and diacritics — KitabiAI handles both natively, then breaks books into structured, hierarchical chapters ready for RAG and semantic search. Bilingual (Arabic/English) UI, sub-10-second processing per book, SEO-optimized output. Built for authors, universities, and libraries sitting on Arabic archives nobody can search yet.
Hey Product Hunt 👋
Three months ago, I watched a podcast where an Arabic author shared his dream: converting his 36 books into searchable web content to enrich Arabic resources online. That stuck with me — Arabic is spoken by 400M+ people, but it's still deeply underserved in AI and search. Most OCR tools mangle Arabic's right-to-left script and diacritics, and there was no simple way for authors, universities, or libraries to make their Arabic archives discoverable.
So I built KitabiAI. Upload an Arabic PDF, and it runs through OCR built specifically to handle Arabic (RTL + diacritics preserved), extracts the table of contents, and generates a clean, structured, SEO-optimized web page.
Each book also comes with its own bilingual RAG chatbot — ask questions in Arabic or English and get answers pulled directly from the book, with citations back to the source.
What's next: Phase 3, where I'm building multi-user management for authors and institutions, extending support to more complex book types like math and science texts, and continuing to improve extraction quality and chatbot retrieval accuracy.
I'd love your feedback, especially if you work with archives, publishing, or multilingual content. And if you know an author or university sitting on a stack of unsearchable Arabic PDFs, I'd love to connect with them 🙏