MakeMyItenary
Team consisting of two WaveMaker engineers specializing in React internals, runtime engineering, and full-stack AI integrations using OpenAI, Node.js, and proxy-based state management.
YouTube Video
Project Description
MakeMyItenary is an interactive travel companion built entirely around Generative UI, transcending static layouts and text-based chats. Instead of returning plain text, our intelligent agent dynamically “paints” every screen the user interacts with.
We utilized the AG-UI pattern and Model Context Protocol (MCP) in a Next.js (App Router) environment, driven by the Anthropic SDK ( claude-haiku-4-5 ). The agent is constrained to a single specialized tool, render_view , which it uses to stream Server-Sent Events (SSE) representing UI states. These events map seamlessly to interactive React components on the frontend. This allows the agent to iteratively refine the travel experience—rendering live Typeform-style follow-up questionnaires, interactive node-graph itineraries ( ItineraryGraph ), and location pickers ( MapPicker )—all while dynamically adjusting the page’s background and mood label to reflect the current context. By treating UI components as the agent’s native output language, we deliver a highly functional, personalized, and visually immersive user experience.