Trippian | Devpost

Inspiration

Planning a trip should be exciting—not overwhelming. But between booking flights, finding hotels, scheduling Ubers, managing food stops, and handling delays, the details can quickly become a burden. That’s why we created Trippian, your personal AI travel assistant. It anticipates and automates every step of your journey so you can stop stressing over logistics and start enjoying the experience.

What It Does

Trippian is an AI-powered travel companion that uses autonomous agents to plan, monitor, and automate your entire journey—from booking flights and tracking delays to scheduling Ubers and suggesting food and lodging. It communicates with you conversationally and acts on your behalf in real time.

How We Built It

Trippian is an intelligent travel-planning app powered by a Python backend (Flask) and a React Native frontend (Expo, TypeScript).
It uses a multi-agent AI system—Calendar, Flights, Hotels, and Travel Assistant—that work together via structured handoffs and a shared ItineraryState.

We integrated:

Google Calendar (availability checks)
Amadeus API (flight & hotel searches)
OpenAI API (natural language understanding and agent automation)

Other technologies include:

Passlib, and Google OAuth for auth
AsyncStorage, React Navigation, and Moti for mobile UX

Just describe your trip, and Trippian handles the rest—from planning to pricing—in one seamless experience.

Challenges We Ran Into

One of our biggest challenges was building truly autonomous AI agents that could collaborate intelligently—not just chained workflows disguised as agents.
We wanted each Trippian agent to:

Have memory
Reason within its domain
Operate independently

Designing the handoff logic between agents while maintaining a shared state took multiple iterations.
We also faced challenges structuring a backend that exposed clean, modular API endpoints for seamless frontend integration.

Accomplishments That We're Proud Of

A clean, intuitive frontend UI that makes trip planning effortless
Successfully implemented AI agents that collaborate intelligently
Built a modular, scalable architecture that’s easy to extend and maintain

What We Learned

Building real AI agents is tough—especially when they need to communicate, share memory, and stay autonomous.
We realized this is as much a design challenge as it is a technical one, requiring thoughtful architecture and experimentation to avoid turning agents into glorified function calls.