Here is what they built, how they built it, and who they are.
Top Prize · 1st Place
StandByStay
A hotel concierge for standby travelers: research first, book only when your seat clears.
VB
Valentina Borovaya
Digital product manager · Aer Lingus
Product summary
Airline employees can fly standby at steep discounts, but their seats are often confirmed at the gate.
That makes booking a hotel in advance nearly impossible. StandByStay is a conversational hotel concierge built for that uncertainty. It searches every possible destination at once, holds a top pick and a backup in each city, and only starts checkout after the traveler confirms which flight they cleared. Nothing is booked until the boarding outcome is known.
Under the hood
Built with Claude Code; the Anthropic API powers the conversation and connects to RouteStack behind the scenes.
Node/TypeScript (Hono) backend with a vanilla JS frontend; hotels render as boarding pass cards to echo the travel context.
On confirmation, the winning card links straight to RouteStack checkout. Sessions are saved, so a booking can resume where it left off.
Maker profile
Valentina is a digital product manager at Aer Lingus and a former Googler who has built B2C products in travel and education for over 15 years. She mentors early-stage founders on product strategy and growth. The Build Challenge was her first hackathon, and StandByStay solves a problem she faces herself as an airline employee.
A family hotel concierge that turns hundreds of guest reviews into one honest recommendation.
VJ
Varun Jain
Travel technology entrepreneur
Finding the right hotel for a family is slow work.
Parents can spend 30 to 40 hours piecing together reviews, maps, and video walkthroughs, and a hotel that suits a couple can still be wrong for a family of six with a toddler. HotelZippo replaces that with a single conversation. An AI concierge learns who is traveling, reads hundreds of guest reviews per hotel, and returns one confident recommendation. If a hotel has a real problem, such as an active refurbishment, the warning appears in red above everything else. Then it hands the traveler off to checkout.
Under the hood
Agentic conversation loop on the Vercel AI SDK, with Claude handling tool calls and turning raw reviews into family-specific intelligence.
Live on Vercel and Supabase with a real data pipeline: guest reviews via Apify, place resolution via Google Places, synthesis via Claude.
Booking runs through the RouteStack hotel API, with a deep-link handoff to checkout and webhooks that track each order from pending to confirmed.
Maker profile
Varun is a travel technology entrepreneur with more than 20 years of experience building and scaling travel and tech products. He joined the Build Challenge to get hands-on with MCP and agentic product building, for his own learning and his own travel needs.
A group trip planner where an AI helps everyone agree on dates, budgets, and destinations.
SR
Stijn Risseeuw
CS & Engineering · TU Delft
Planning a group trip usually means messy chats, scattered spreadsheets, and conflicting preferences.
MatchRoute brings everyone into one shared dashboard where an AI assistant helps the group negotiate budgets, dates, and destinations. Interactive maps, calendars, budget meters, and voting polls appear right inside the trip chat, and a conflict timeline helps settle differences before anything is planned for real. Once the group aligns, MatchRoute generates five tailored trip recommendations that match the group's collective profile.
Under the hood
Next.js and TypeScript with a Neon Postgres database; Google Gemini via the Vercel AI SDK drives generative UI elements such as Leaflet maps, calendars, and polls.
An OpenAPI client generated from RouteStack's spec pulls live hotel, flight, and rental car data.
A cross-device prompt queue and a dedicated conflict timeline page keep multi-user planning in sync.
Maker profile
Stijn is an Honours Computer Science and Engineering student at TU Delft who teaches as a university TA and builds full-stack platforms. He gravitates toward complex algorithmic problems under pressure, which makes hackathons a natural fit.
