Travel Planning AI Agent with MongoDB Atlas, Gemini & Vector Search

This workflow demonstrates an advanced agentic AI setup for travel planning, featuring two interconnected flows: an ingestion pipeline and an interactive AI agent. The ingestion flow receives documents via webhook, embeds the title and description using OpenAI (or compatible provider), and stores them in MongoDB Atlas Vector Search within the 'points_of_interest' collection. This enables efficient retrieval of relevant travel data like locations and attractions.

The core AI agent flow activates on chat messages, leveraging MongoDB Atlas for persistent chat memory and a Vector Search tool for context retrieval. Powered by Gemini LLM, the agent dynamically queries embeddings to provide informed responses on travel queries, such as recommendations or details about points of interest. This reduces custom wiring overhead, streamlining memory management, similarity search, and orchestration.

Benefits include scalable long-term memory, RAG-enhanced accuracy, and seamless integration of native n8n nodes for MongoDB Atlas. Use cases span travel apps, virtual assistants, customer support bots, or any domain needing vector-based retrieval with conversational AI. Prerequisites: MongoDB Atlas cluster, OpenAI/Gemini API keys.