Team: Latton Lab
Team Member: Chang Cheng Jun, Kelvin Ng Yi Soon, Lee Yun Sheng, Stephen Sii Zhong Chee
Problem Statement: Travel Planner
Video Presentation: [Latton Lab]
Presentation Slides: [Latton Lab Slide]
Application available on market lacks functionality, or lack multi functionality that can be put together to provide an all in one user experience in planning their own travel. Besides, Boring UI, especially too many similar vibe coded app with similar UI layout, which makes travel planning feel dull and boring, at the same time having a weak multi user load handling and backend computation structure.
User are the people that can feel straight forward difference between a useful and handy app or an ordinary travel planning application with dull functions which everybody can make one.
First, available application on market such as Booking.com, Trivago, Traveloka are some most popular market products. However, they all have a same huge problem — No group planning feature. We don’t mean that those apps are not usable, but each users can only plan for their own trip, making preference synching between different users lack of collaborative and interactive elements.
We added features as below:
- Generative UI (Make sure each and every user’s experience are unique, just like a game!)
- Group Planning Mode (Generative UI generate different location recommendations for each group user, where each of them pick their personal favorite spots, and team voting for the best, optimal location.)
- AI based itinerary planner (User pick their location, AI do the rest, including location research, rating, things to do, best routes, all based on user’s budget and personal preference)
- User preference tracking system (Let AI understand each user better, from several perspectives including food, activeness, and so much more, ultimately providing perfectly tailored, completely personalized travel plan)
| Idea | Why it was dropped / kept |
|---|---|
| Dual Mode Support (Solo & Group Travel) (Chosen) | Allows the platform to serve both individual travelers and collaborative group trips through tailored flows. |
| Generative UI with Animated View (Chosen) | Replaces tedious manual searches with engaging visual prompts, enabling users to indicate interest by clicking animated view. |
| Voting & Highest-Vote Aggregation (Chosen) | Automatically tallies group members' choices and selects top-voted destinations and hotels to prevent decision deadlocks |
| Mandatory User Account Registration (Dropped) | Replaced by direct access via temporary room codes and display names to reduce onboarding drop-off. |
| All-in-One PDF Travel Report Output (Chosen) | Summarizes all travel destinations, timing, and hotel details into a single printable document |
| Two-Stage User Onboarding (Chosen) | Collect core parameters (destination, date, budget, team size), then collect personality traits. Eg: -How much of a foodie are you? -How active are you? -Do you like nature? |
| Built-In Real-Time Group Chatroom & Messaging (Dropped) | Dropped to prevent messy UI view; users communicate directly through link sharing and vote choices. |
| One-Tap Like / Wishlist Favorite Action (Chosen) | Allows users to quickly save places to their personal preset favorites via a heart button. |
| Interactive Destination Detailed View (Chosen) | A detail modal over a stylized 3D scene showing showing breadcrumbs, favorite/share buttons, photos, and contextual summaries of destination will be shown when user clicks on a destination |
| AI Voice-Activated Search Interface (Dropped) | Replaced by visual animated cards and direct click interactions for faster browsing. |
| Passport & International Visa Requirement Checker (Dropped) | Omitted to keep onboarding fast and free of complex legal/government API integrations. |
| PC Web Page (Dropped) | Discarded to ensure a flexible web app which is accessible on mobile devices on the go. |
| Location Header Navigation (Chosen) | Displays clear hierarchical paths (e.g., Country / City / Landmark) to keep users oriented within generative card flows. |
| Interactive Hotel Choosing Process (Chosen) | Gives users control over hotel options via interactive elements. |
| Direct Invite & Social Link Sharing (Chosen) | Includes a quick-share action (paper plane icon) to easily send room invites and destinations to desired partners. |
| Saved locationSidebar (Chosen) | Keeps favourable destinations accessible throughout the selection stage (❤️icon). |
The rough sketch maps out the core app structure, including various screens and a landing page layout, alongside standard navigation elements like an address bar. Instead of standard booking features, it envisions a generative UI where users can research and explore destinations through an interactive, game-like experience.
The flowchart outlines a user journey featuring both "Solo" and "Collaborative" planning paths. The process begins with "Room creation" and proceeds through a two-stage user onboarding phase that gathers trip details—such as destination, date, and budget—alongside personality checks and user preferences. The flow then transitions into a "Generative UI" stage for research and content generation, which allows collaborative users to vote on options like places and hotels. Ultimately, the journey concludes at "Stage 2: Finalized report," which is generated either from personal picks for solo users or by recording the highest-voted choices collected from everyone in a collaborative group.
| Date | Mentor | Feedback Received | What Was Changed |
|---|---|---|---|
| 08/09/2026 | Janelle Tan | The flow is very clear, users can know what to do when entering the platform. The whole solution is user-friendly. Very memorable. | Add more detail for the location and travel plan in the prototype. |
UI Prototype: [Figma Prototype Design]
 Mode Selection Where the user picks Solo Mode (plan alone) or Collaborative Mode (plan with friends in real time). |
 Trip Details Collects destination info — country, state, exact place — before generation kicks off. |
 Room Creation Collaborators join via a 4-digit room code; the host enters the shared destination for the group. |
 Generated City Nano Banana generates a 45° isometric miniature city scene with landmarks and current weather baked into the lighting — no icons or text needed. |
 Landmark Insights Clicking a landmark has the backend identify it and generate travel-planner-style insights, shown alongside the landmark. |
 Saved Locations Hearting a landmark adds it to the trip plan in the slide-up card, building the itinerary as the user explores. |
 Export Travel Plan AI automatically sums up the itinerary, handling the research and planning so everything's ready to go. |
 Plan Ready The finished plan exports to a detailed PDF — in Collaborative Mode, only the room host can trigger the export. |
| Existing Solutions | Novel Features |
|---|---|
| Poor UI/UX, which causing eye strain and confusion of user experience. | Applying the generative UI by using the LLM to generate the dedicated city overview dynamically, user can choose the desired location by just clicking the desired location. |
| Not support group trips planning, need the user to manually mediate each other, cannot make a collective decision. | Allows real-time collaboration, which generating the dedicated travel plans by creating a specialised group for them, referring each users' preferences and users polling. |
| Dull user experience, lack of creativity. Unable to catch users eyes. | Applying the flashcard voting to provide gamification experience, which allows multiple user to vote their preferred destination simultaneously for deciding their group travel destination. |
| Lack of customization, travel plan hard to meet with users' anticipation. | Flexibly suggest the optimized route, destination, transport and hostel based on the users preference, such as budget, types of travel, preference score for foodie, nature, shopping and active. |
| Function deficiency, most of the user unable to support single task. | Integrate destination suggestion, route planning, transport planning, flight ticket and hotel booking by utilizing the AI and external API. |
| Travel plan are not customizable for difference seasons, causing the users to missed with the seasonal special activities. | Dedicated travel plans based on the specified seasons, ensure all of the activity is available during that seasons. (E.g. Skiing for winter seasons, Visiting NYC Time Square during Christmas) |
| Description of the travel attractions are outdated, need to updated manually the latest information. | AI-generated description, it can capture and the real-time information from internet, which provide the most accurate information to user. |
Tech stack
| Layer / Component | Technology Selected | Reason |
|---|---|---|
| Frontend | HTML5, CSS3, Vanilla JavaScript | Removes heavy framework overhead for fast rendering; gives fine-grained control over DOM to enable a custom Neumorphic tactile design system and super fast interactive canvas manipulation. |
| Backend | Google Cloud Run | Auto-scaling containerized execution, handling compute-heavy image parsing, AI coordination, and PDF generation |
| Generative UI / Generative Surface | AI generative canvas | AI-generated interactive visual scenes |
| AI Vision & Persona Engine (APIs) | Google Gemini 3.5 Flash / 3.1 Pro & Nano Banana Pro | LLM Model multimodal coordinate grounding to semantically map unstructured pixel clicks (x, y) coordination to real-world landmarks and quick structured JSON output for the travel persona insights. Nano Banana will be pecialized in rapid rendering of high-resolution (1920 * 1080) 45° isometric 3D miniature scenes. |
| Backend | Python (FastAPI) / Rust | Native, low-latency SDK integrations for generative AI models |
| Database Integration | Supabase (PostgreSQL + Realtime Engine) | Free and out-of-the-box websocket channels to validate group room codes, aggregate live votes, and keep up with wishlists, all without running dedicated socket servers. |
| Frontend Deployment | Cloudflare Pages | CDN network with instant loading speeds, low latency for static assets, and automated CI/CD directly from Git. |
| Platform | Progressive Web App (PWA) | Enabled via manifest.json and service worker caching |
| Compactability | Android & IOS | Mobile installation to the home screen without app store review delays. |
Scan report · 2026-10-13
- ✓ Prohibited terms or links
- ✓ Repository eligibility
- ✓ slopscore.md paperwork
- ✓ Content policy
- ✓ Risk review — +15 no detected language (no code?)
From the balcony · 0 of 4 clapped
Schnitzel, Cap'm Slop, Princess and Crusoe read it and passed. Their reasons are on the balcony, with every other verdict.
Critics are accounts on this site with no GitHub account behind them. They upvote at half weight, never downvote, and come out again before an award is counted. Who they are.
0 comments
log in to comment.