MUMBAI, India, Jan. 2 -- Intellectual Property India has published a patent application (202541125179 A) filed by Keshav Memorial Institute Of Technology, Hyderabad, Telangana, on Dec. 11, 2025, for 'docgenie - intelligent document question answering system.'

Inventor(s) include Ms. Madhurika Budaraju; Mr. K Rajesh Kumar; Mr. Para Upendar; Ms. P. Snigdha; Mr. B. Charan Sai Reddy; Ms. D. Anushka Reddy; Ms. K. Harshitha Reddy; Mr. K. Manoj; and Ms. I. Shruthi.

The application for the patent was published on Jan. 2, under issue no. 01/2026.

According to the abstract released by the Intellectual Property India: "This project focuses on building a smart Question-and-Answer (Q&A) system that can read documents and answer user questions based on the content in those documents just like asking a knowledgeable assistant. The system supports different types of documents such as PDFs, Word files, text files, web pages, and notes. Once these documents are uploaded, the system breaks them into meaningful parts and converts them into semantic embeddings, which means it stores the meaning of the content, not just the words. These are saved in a special type of searchable storage called a vector database. When a user asks a question, the system first finds the most relevant parts of the documents and then uses Google Gemini, a powerful AI model, to understand the question and generate a clear, accurate answer. The system also uses a technique called Retrieval-Augmented Generation (RAG) which means it combines document search with intelligent answer generation. A key feature of this system is memory. It remembers what the user asked before (short-term memory) and can even remember past conversations (long-term memory), which helps it give more personalized and meaningful answers over time. The system also learns and improves continuously by collecting feedback, correcting errors, and testing different answer styles. It is tested using standard question sets like SQuAD 2.0 and CoQA to make sure it's accurate, fast, and reliable. The project is built using FastAPI or Flask for managing the system's backend and Streamlit or React to design the user interface. Overall, this system transforms the boring task of document search into a smart, helpful, and interactive experience, useful for students, teachers, researchers, and anyone who works with a lot of documents."

Disclaimer: Curated by HT Syndication.