-
-
travelfix
The Problem
Travel disruptions can turn a simple journey into a stressful chain of problems. A cancelled or delayed flight can mean finding a hotel, transportation, alternative routes, and contacting providers often while the traveler is already under pressure.
TravelFix was built around a simple idea:
When travel goes wrong, the traveler shouldn't have to solve everything manually.
What is TravelFix?
TravelFix is an AI powered travel recovery agent that detects disruptions, analyzes the situation, finds recovery alternatives, compares available options, and guides the traveler toward the best recovery plan.
The experience is designed as a step-by-step recovery workflow:
- Detect — identify the travel disruption.
- Analyze — understand the consequences and constraints.
- Find Alternatives — discover possible recovery options.
- CALL-E Calls — use the CALL-E voice agent layer for provider communication.
- Compare Options — evaluate the available recovery choices.
- AI Recommendation — recommend the most suitable option.
- Traveler Approval — let the traveler approve the recovery before finalizing it.
Why We Built It
The inspiration came from imagining the moment when a traveler receives a cancellation notification and suddenly has to coordinate several different services.
Instead of giving the traveler another checklist, TravelFix turns the disruption into an agent-driven recovery process.
The goal is not simply to provide information. It is to help transform:
Travel disruption → recovery options → informed decision → recovered journey
How We Built It
TravelFix uses a Flutter web frontend and a Python FastAPI backend.
The frontend provides the interactive recovery experience and visualizes each stage of the recovery process.
The backend handles the recovery workflow and integrates with the CALL-E API for the voice-agent portion.
CALL-E is used as the communication layer for contacting travel providers. For the submitted demonstration, the CALL-E interaction is presented in simulation mode, so the demo does not place real phone calls.
What We Learned
Building TravelFix taught us that an AI travel agent needs more than a conversational interface. It needs a structured workflow that understands the disruption, keeps track of constraints, evaluates alternatives, and gives the traveler control over the final decision.
We also learned how important it is to design graceful fallbacks when external services are unavailable or asynchronous.
Challenges
One of the biggest challenges was connecting an interactive frontend recovery workflow to an external voice-agent API while keeping the experience reliable.
External calls can be asynchronous and may take time to progress. We therefore designed TravelFix so that the recovery experience can continue using simulated provider interactions when a live voice call is not appropriate for the demonstration.
Another challenge was keeping sensitive API credentials server-side. The CALL-E API key is stored in the backend environment and is not exposed to the Flutter client or committed to the public repository.
The Result
TravelFix demonstrates how an AI agent can take a stressful travel disruption and turn it into a structured recovery experience.
Instead of asking:
"What do I do now?"
TravelFix aims to give the traveler:
"Here are your recovery options. Here's what we recommend. You decide."
Log in or sign up for Devpost to join the conversation.