TripSwipe
Tagline: Swipe together. Plan a trip everyone says yes to.
Product tagline: Swipe through AI-curated experiences and turn your group's travel mood board into a day-by-day itinerary.
Inspiration
Travel is what makes me feel most alive, and it has always been something I experience with other people: family, friends, and whoever is ready for an adventure.
I flew from Seattle to Los Angeles for LA Hacks, my first time in California, and wanted to make the most of the days before the hackathon began. I started with Google, Expedia, and Reddit, but I ran into the same problem: these platforms assume you already know what you want before you can meaningfully discover it. Then I tried ChatGPT, which gave me a premade itinerary that did not reflect what I actually wanted to do.
Meanwhile, the inspiration that makes us want to travel lives on Instagram, Pinterest, and saved photos. But that inspiration rarely makes it into a plan everyone agrees on.
I wanted to bridge that gap. What if planning a trip started with the feeling you wanted to chase, rather than a list of places you already knew? What if you could turn your inspiration into experiences, swipe through them with your travel group, and build an itinerary together?
That idea became TripSwipe, a collaborative, AI-powered travel planner that transforms inspiration into a trip your whole group is excited about.
What it does
TripSwipe combines visual discovery, AI-powered personalization, and group decision-making into one end-to-end travel planning experience.
Set up the trip. Start with a destination, dates, travelers, budget, and the overall vibe you are looking for.
Build a visual mood board. Upload inspiration photos that capture your ideal trip, from mountain escapes to beachside cafes. AI analyzes the images to identify visual themes and preferences that guide recommendations.
Discover experiences through swiping. TripSwipe generates a personalized deck of food spots, iconic attractions, local experiences, and hidden gems based on your inspiration, interests, travel dates, and budget. Swipe right to save, left to skip, and mark must-dos.
Make decisions together. Invite friends and family by email. Everyone can swipe, upvote, and downvote experiences, with group preferences brought together on a shared leaderboard. Instead of losing decisions in a 40-message group chat, everyone gets a voice in the plan.
Turn the group's choices into an itinerary. Explore relaxed, packed, and budget-conscious itinerary options side by side, with explanations of the trade-offs. Then generate an editable, day-by-day schedule that evolves as the group finalizes its preferences and budget.
The result is a journey from "What do we feel like doing?" to "This is the trip we're taking."
How I built it
I independently built TripSwipe from concept to implementation, using Base44 for the entire application. I handled the user research, product design, frontend, backend logic, AI integrations, data modeling, and end-to-end implementation myself.
The frontend uses React and Tailwind, while Base44 entities manage trips, mood-board photos, experiences, saved picks, votes, messages, invitations, and itinerary days. Backend functions power the AI and collaboration workflows, including visual preference extraction, experience generation, itinerary generation, trade-off explanations, invitations, voting, and trip chat.
A key design decision was treating AI as a structured generation system rather than simply adding a chatbot. Each generation follows a defined schema and is validated before its results reach the interface. This keeps experience cards and itinerary options consistent and usable instead of relying on free-form responses that may be difficult to render or act on.
Before writing code, I surveyed more than 60 people at the hackathon and beyond to understand whether the problem extended beyond my own experience. I then mapped the user journey through vision boarding and Figma before building the application end to end.
Challenges I ran into
- Making AI recommendations trustworthy.
AI-generated recommendations can contain fabricated places, broken links, or unrealistic costs and durations. I implemented validation to check required fields, map links, reasonable cost and duration values, and supported experience categories. Malformed recommendations are discarded rather than displayed.
- Turning inspiration into personalization.
A mood-board image communicates a feeling, not necessarily a structured preference. I needed to translate visual inspiration into useful taste tags and combine them with traveler interests, trip dates, and budget to generate relevant recommendations.
- Designing for group consensus.
Collaborative planning requires more than sharing a list. Swipes, votes, and messages must remain associated with the people who created them, while trip membership determines who can access or modify a plan. I implemented server-side checks to enforce these boundaries.
- Balancing AI cost, performance, and reliability.
I introduced per-person daily generation limits, rate-limited invitations, and reused existing experience decks instead of regenerating them unnecessarily. These safeguards help control usage while making the experience faster and more consistent.
- Building and shipping the entire product independently.
As the sole builder, I was responsible for connecting every part of the application, from generating experiences to collecting group votes and producing a usable itinerary. I had to prioritize carefully, troubleshoot across the stack, and make sure the core user journey worked as a complete product rather than a collection of disconnected features.
Accomplishments I'm proud of
Building a complete product independently. I took TripSwipe from initial research and design through full-stack implementation, integrating AI-powered discovery, collaborative voting, and itinerary generation into one working application.
Using AI to support real decisions. Visual analysis informs personalized experience generation, while structured itinerary generation presents different planning options and explains their trade-offs.
Solving for group consensus. Shared mood boards, swiping, voting, and chat address the coordination problem at the heart of planning trips together.
Building with reliability in mind. Schema validation, access controls, generation limits, and reuse of existing recommendations help make the application more reliable and efficient.
Grounding the product in user research. More than 60 survey responses helped me investigate the problem and inform the planning experience before implementation.
Creating a distinctive product identity. I designed the travel-journal aesthetic with warm paper tones and hand-drawn illustrations to make planning feel like the beginning of an adventure rather than another task on a checklist.
I'm especially proud that I was able to bring the entire experience together independently on Base44, from the initial idea and research to the working application.
What I learned
One of my biggest lessons was that building a better product often means removing complexity, not adding features. Iteration helped me focus on the features that move users from inspiration to a decision instead of building functionality simply because it was possible.
I also learned that AI works best when given clear constraints, structured outputs, and validation. A compelling recommendation is only valuable if the application can trust it and turn it into something actionable.
Conducting more than 60 surveys helped me challenge my assumptions and approach the problem from the user's perspective. It reinforced the importance of gathering feedback early, testing the actual workflow, and making product decisions based on observed needs rather than personal preferences alone.
Most importantly, I learned that travel planning is inherently social, but the tools we use to plan it are often individual. TripSwipe explores what happens when the planning experience is collaborative from the start.
What's next for TripSwipe
The next step is to make the journey from inspiration to itinerary even more seamless.
Import inspiration from Instagram Reels. Allow users to share a Reel with TripSwipe and extract the featured destination or experience so it can be explored, swiped on, and saved to a trip.
Export to Google Calendar. Turn the finalized itinerary into calendar events so the plan can move seamlessly from the app into everyday life.
Make recommendations actionable. Explore integrations with travel companies so users can discover, save, and eventually book relevant experiences directly from their plans.
Improve through continuous user research. Conduct more interviews, observe how groups use the product together, and refine the onboarding, swiping, voting, and itinerary flows based on real feedback.
Ultimately, I want TripSwipe to become the place where a group's scattered travel inspiration turns into a shared plan, and where deciding what to do together becomes part of the excitement of going somewhere new.
(I dont have a video submission!)
Built With
- agents
- base44
- chatgpt
- claude
- figma