Trip Planner Agent

Your trip start here!

Inspiration

We were tired of jumping between websites—checking flight prices on one, hotel availability on another, attraction opening hours on a third, and weather forecasts somewhere else. We set out to build something that could consolidate trip planning into one intelligent conversation, provide accurate real-time information, create customized itineraries, and most importantly, save travelers hours of research.

What it does

Our Smart Trip Planning Agent transforms the frustrating experience of planning a trip into a seamless, intelligent conversation. When you ask for a trip plan, the system works behind the scenes through a sophisticated three-step process: first, it creates a detailed day-by-day itinerary using GPT-4's planning capabilities; then, it automatically searches the web for real-time information about opening hours, current ticket prices, weather conditions, and local events using the Tavily API; finally, it synthesizes everything into a comprehensive, actionable travel plan that includes current prices, practical tips, and direct booking links. Instead of spending hours searching across multiple websites, users get everything they need in one intelligent conversation - from the initial itinerary to current pricing and availability, all presented in a clean, easy-to-follow format that makes trip planning simple and reliable.

How we built it

We structured our system in three key phases to ensure a smooth and intelligent trip planning experience. First, we built a strong architectural foundation using a stateful AI framework that supports multi-step workflows. This allowed us to design an agent capable of managing complex interactions while keeping user data organized and consistent throughout the planning process.

The agent operates in a sequential workflow: it first generates an initial trip itinerary based on user input, then gathers real-time information such as attraction hours and weather, and finally combines everything into a complete travel plan. By integrating a live search system, the agent ensures that users receive accurate and up-to-date trip details tailored to their preferences.

Challenges we ran into

One of the biggest challenges we faced was managing the volume and organization of information generated by the AI. Initially, the search results and itinerary details returned by the model were overly long, repetitive, or disorganized, making them difficult to structure into a clear and usable travel plan. We found that even small changes in prompts could dramatically affect the output, which led us to go through multiple iterations of prompt engineering and workflow adjustments. It took a lot of trial and error to strike the right balance—refining how the AI generates and processes information so that the final output would be both concise and meaningful for users.

What we learned

Through this project, we learned the importance of having a clear idea of what we wanted to build and how to approach it before writing any code. Taking time to plan helped us define our goals, understand the problem more deeply, and stay focused throughout the development process. This upfront clarity made our workflow more efficient and helped us avoid unnecessary confusion or rework later on.