Final Year Project

Project Title

Zakheera-e-Khabar

Description

Zakheera-e-Khabar is an Urdu newspaper archival platform that digitizes the newspapers, making them easy to search, read, and explore. This project helps preserve print media by converting scanned newspaper pages into searchable text using high-accuracy Urdu OCR, AI-powered summarization, and semantic search functionality. This project makes use of tools like Gemini 2.0 Flash for OCR, ChromaDB for semantic search with dense vector embeddings, and an mBART-based model trained specially for Urdu summarization. Built with Python, React.js, MongoDB, and Flask, the platform is user-friendly and scalable. With features like collections and trend analysis, it helps users discover historical events, follow changes in public opinion, and explore editorial narratives. It also offers filters by date, newspaper, and topic to help narrow down results.