Published Paper Details:

In this paper, we present an AI-powered Library Chatbot designed to provide personalized book recommendations and enhance the accessibility of library resources. Students often face difficulties in locating the right books in college libraries due to the vast number of available titles across diverse subjects. For instance, a student preparing for UPSC or CA examinations may struggle to find the most suitable book based on their level of understanding or area of focus. To address this challenge, we developed a Retrieval-Augmented Generation (RAG)-based system, built using the LangChain framework, that efficiently provides detailed information such as book accession number, author, edition, subject, language, place of publication, and rack location-- significantly reducing search time from nearly 20 minutes to a few seconds. The chatbot is built using Python, Streamlit, HTML, and CSS for the user interface, integrated with ChromaDB as the vector database and the Hugging Face AllMiniLM model for word embeddings. The GROQ API with Google 2B LLM serves as the core language model. The system was trained on a dataset of 26,000 books provided by the college librarian and is deployed on Streamlit Community Cloud for public access.