Inspiration
We were inspired by the idea of enterprise ai travel. Through our internships we've always had to book things ourselves for company trips, so the idea of automating this is something that we thought would be useful for us and many other people.
What it does
It uses multiple agents in order to travel, book hotels, and negotiate to find the cheapest prices. It also allows managers to add employees onto their business trips and plan itinerary. This information is then sent as a notification to employees in the form of tickets, budget and an itinerary list.
How we built it
I built WAYPORT as an agentic travel platform where AI agents—not static forms—plan, price, and coordinate entire trips. The frontend is purely on jaclang with jac hammer with Framer Motion, and Three.js for a cinematic feel, while Mapbox powers an interactive "Map OS" that visualizes every stop on the itinerary. Trip creation starts with a Concierge intake that uses Gemini to turn natural language into structured trip data, which an Orchestrator then routes to specialized agents (Planner, Flights, Hotels, Financial, and others) that each log their actions for full auditability under a risk-tiered policy engine. The Planner builds day-by-day itineraries with real pricing, and a scoring system (Rove) compares payment options—cash, miles, points—to optimize cost. Third-party APIs (Amadeus for flights, Stay22 for hotels, OpenWeather, Clerk for auth) round out the system. It also is connected to a mobile app where travelors on a team can access trip information and boarding passes.The result is a multi-agent, map-visualized operating system for planning trips solo or coordinating group travel at scale.
Challenges we ran into
Biggest challenge was setting up the agentic orchestration. It was really annoying to put multiple at once.
Accomplishments that we're proud of
We're proud of the fact we accomplished a CRAZY amount of stuff in a short time. We worked very efficiently and I'm happy we were able to come up with a functioning app with agent orchestration, and database.
What we learned
We learned a lot about agent orchestration with pydantic and it was really cool to be able to. implement payments with rain's stable coin platform.
What's next for Wayport
Better authentication system, and adding more features that allow us to connect with an enterprise
Log in or sign up for Devpost to join the conversation.