Inspiration
Planning a trip usually means jumping between multiple tools for flights, destinations, and itineraries. We wanted to make that experience more natural by letting a user describe their travel needs in one request and have an AI agent coordinate the planning process behind the scenes.
Voyage was built around the idea of using MCP (Model Context Protocol) to orchestrate specialized tools and turn a simple travel request into a structured trip plan.
What it does
Voyage is an AI-powered trip planning assistant.
A user can provide a request such as their destination, dates, preferences, and travel requirements. The application then processes the request and coordinates the available travel tools to generate a trip plan.
The project demonstrates MCP-based orchestration, connecting the AI agent with travel-related functionality such as flight search and itinerary planning.
How we built it
The project was built with Python and Flask, with MCP used as the orchestration layer.
The system connects the trip-planning agent with external capabilities, including the Duffel API for flight-related functionality. The project was also designed with Alexa+ in mind, allowing the same planning experience to be exposed through a conversational interface.
The overall flow is:
- The user submits a natural-language trip request.
- The agent interprets the travel requirements.
- MCP coordinates the appropriate tools.
- Travel information is retrieved through the connected APIs/tools.
- The information is combined into a structured itinerary.
- The resulting trip plan is presented to the user.
Challenges
One of the main challenges was coordinating multiple components while keeping the user experience simple. The user should only need to describe what they want to do, while the orchestration layer handles the underlying tool calls.
Another challenge was connecting the AI agent with external travel functionality in a reliable and structured way. MCP helped provide a consistent interface between the agent and the tools it needed to use.
Accomplishments
We built a working end-to-end prototype that demonstrates AI-driven trip planning with MCP orchestration.
The project brings together a conversational interface, an AI planning workflow, external travel APIs, and a generated itinerary into a single experience.
What's next
We would like to expand Voyage with more travel providers and richer itinerary planning, including hotels, activities, transportation, and personalized recommendations.
Future versions could also improve conversational follow-up, allowing users to modify parts of their trip without having to start the planning process again.
Demo and Source Code
GitHub: https://github.com/Shriya-JB/Alexa-trip-planner
A short demo video is included with the submission to show the trip-planning workflow and MCP orchestration in action.
Log in or sign up for Devpost to join the conversation.