Trip Tailor: AI-Powered Travel Planning with Itinerary Generation and Chatbot Assistance

Abstract

AI-driven travel planning systems, such as those using ChatGPT and content-based recommendation engines have shown promise in itinerary generation but often lack real-time adaptability and retrieval-augmented mechanisms. This limits their ability to provide dynamic, context-aware recommendations. The proposed work introduces an AI-powered system that streamlines travel planning with tailored itinerary generation and real-time travel assistance. The itinerary generator, powered by Google Generative AI (Gemini API), creates custom travel plans based on user inputs like destination, duration, budget, and companions. Integrated with the Google Places API, it provides detailed recommendations for hotels and day-by-day itineraries, including optimal timings, travel durations, and ticket pricing. The travel assistant chatbot, utilizing a RAG pipeline with the Deep Seek-R1 model, offers real-time, conversational insights through Wikipedia API data. FAISS-based indexing ensures fast, accurate responses, supported by pre-processing techniques like text normalization and entity extraction. This dual-component system delivers a scalable and user-friendly solution, combining generative AI with robust retrieval mechanisms to meet the growing demand for personalized and adaptable travel planning.