An AI Powered Multilingual Guide for Tamil Nadu’s Heritage Tourism using Large Language Models

Cultural heritage tourism greatly aids the preservation and promotion of historical sites, but many tourists struggle to obtain up-to-date, multilingual, and contextualized information. The visitor experience is limited by the lack of personalization, interactivity, and language accessibility in traditional heritage guides and static information boards. With the help of Large Language Models (LLMs), natural language processing (NLP), and geospatial technologies, this paper offers a voice-assisted, AI-powered heritage guide that offers dynamic, multilingual insights about Tamil Nadu's historical sites. Through the integration of an LLM-based interactive tour assistant with text-to-speech and speech-to-text capabilities, visitors can ask conversational questions regarding the significance and history of the site. Real-time data APIs also offer information on visitor trends, the best times to visit, and weather conditions. The suggested framework ensures smooth exploration of heritage sites by using the Google Maps API for navigation. Additionally, adaptive learning is used by the solution to tailor suggestions according to user preferences. Benchmarking techniques are employed to compare the performance with conventional tour guide knowledge and information generated by Large Language Models (LLMs). It is designed to be web-based and compatible with mobile devices, making it convenient for both tourists and culture enthusiasts.