This project presents an advanced travel recommendation system designed to overcome the limitations of traditional keyword-based approaches. It uses semantic analysis, GraphRAG (Graph Reasoning with Augmented Graphs), and large language models (LLMs) to generate personalized and context-aware travel suggestions. The system is built using Python and integrates the Llama-3.2-1b model via Ollama for entity extraction and query processing. Travel data is cleaned, vectorized using the Nomic embedding model, and stored in a Neo4j graph database to enable intelligent multi-agent reasoning, including query refinement and summarization. The system improves recommendation accuracy, highlights lesser-known destinations, and supports sustainable tourism through AI-driven insights.