MCQ Generator
V. Roshan Kumar¹, Chinta DivyaSurathi², Galla Bharghav Ram³, Kola Govardhan⁴, Kakumanu Amani ⁵
1Assistant Professor, Department of CSE(AIML)Bapatla Engineering College, Bapatla 522101, AP, India 2,3,4,5 Student, Department of CSE(AIML)Bapatla Engineering College, Bapatla 522101, AP, India
roshan4linus4550@gmail.com, divyasurathi2704@gmail.com, bharghavramgalla@gmail.com,govardhankola95@gmail.com, kakumanumani@gmail.com
Abstract—The online learning resources made it possible to effectively enhance and showcase these resources; thus, making automated assessment generation even more fundamental in contemporary era education. In this study, we introduce a smart question generation system powered by LLMs to autogenerate different styles of assessment questions from educational material. The proposed system enables users to upload documents or provide website URLs from which MCQs, fill- in-the-blank questions, true-false questions, and short answer queries can be created. With an extensible architecture, the framework combines several LLM providers allowing users to use their own API keys for higher flexibility and performance. Using sophisticated prompt engineering techniques, the models are directed toward generating structured and localized questions. The system covers preprocessing and text extraction modules which use unstructured educational material as inputs to produce appropriate inputs for question generation. A web- based interface designed using Python and Flask allows teachers and students to intuitively interact with the system. The proposed system thus increases adaptability to different educational contexts by supporting diverse types of queries and smoothly integrating various models at run-time. Experimental observations show that the system generates answers that are logical and contextually appropriate at a tiny fraction of the time and effort required to generate questions manually.
Keywords—Prompt engineering, educational technology, web content extraction, multiple choice questions, automated question generation, large language models, personalised learning.
