Fahad0saad/arabic-ocr-v2

feat: unlimited pages + concurrent OCR processing (3 parallel workers)

Fix: correct Qwen model IDs - remove :free suffix, use 8B default (fixes OCR API errors)

Fix PDF export (Amiri fonts, BOM for TXT/MD, media_type)

Add jinja2 dependency

Fix Dockerfile: libgdk-pixbuf-xlib-2.0-0 + remove fonts-noto-arabic

Add README.md with Docker SDK config

Arabic OCR v2 - FastAPI + Docker deploy

initial commit