BauddhaAI is a RAG (Retrieval-Augmented Generation) system that supports research in Buddhist studies. Out of the 14,000 papers in the Journal of Indian and Buddhist Studies (Indogaku Bukkyōgaku Kenkyū), it retrieves the 20 most relevant to a given question and then asks Google Gemini, together with those papers, to produce an answer. The system was first presented at an international symposium held at the University of California, Berkeley in October 2024.
BauddhaAI takes advantage of Gemini's large context window (1 million tokens). Depending on the question, it can provide quite useful answers, and because the answers are grounded in open-access papers, sources are easy to verify and hallucinations are less likely.
The results generated by this system are to be used at the user's own responsibility. This system is developed by Kiyonori Nagasaki (International Institute for Digital Humanities / Faculty of Letters, Keio University), but the developer bears no responsibility for the generated results. For details about the system's architecture, please see the research report below.
Kiyonori Nagasaki, "BauddhaAI: Development of a RAG System for Buddhist Studies Using Generative AI," IPSJ SIG Technical Reports, Humanities and Computers (CH), 2025-CH-137, 8, pp. 1-5.