main page
choose the dates
select activities
adjust the budget
final destination
main page (design 2)
choose the dates (design 2)
select activities (design 2)
swipe to choose
final destination (design 2)

SkyGrouper Trip Planner

SkyGrouper is a Model Context Protocol (MCP) web app where cooperating AI-agents turn a group chat’s “we should totally travel!” into an actionable, book-ready getaway.
It juggles multiple departure airports, personal budgets and shared interests, then delivers the cheapest common destination plus deep-links to book the flights—all through a playful swipe interface.

Key Features

What it does

SkyScanner MCP Server
Shortlisting of candidate cities based on group interests
Concurrent LLM-powered destination overview and flight plan generation
Multi-origin support with individual budgets
MongoDB integration for storing and retrieving trip requests

Inspiration

Group trips often die in the chat: too many spreadsheets, price disagreements, and “who’s free when?” paralysis.
We wanted to shrink days of back-and-forth into a 30-second, single-screen experience that feels more like a game than a form.

How We Built It

Layer	Tech
Front-end	TypeScript
Back-end	FastAPI (Python)
Flight data	Skyscanner Travel API
Data store	MongoDB Atlas

Challenges

Multi-origin price maths – finding a city that’s “fair-priced” for everyone.
Conversational vs. form UX – early chatbot prototypes felt robotic and the classic multistep form was dull; the swipe deck finally nailed engagement.

What We Learned

Good UX can hide a ton of optimisation logic—users never see the date-intersection math or budget filters, they just swipe.
APIs in travel are surprisingly capable;