WanderLens

About the Project

Inspiration

Planning a trip today is messy. You jump between a flight search site, maps, blogs, forums, and hotel pages — and after hours of searching, you often end up more overwhelmed than when you started.

We wanted to build something simpler: a tool that feels like talking to a well-traveled friend who can instantly give you a full plan, show you real prices, and guide you with actual experiences from other travelers.

We were also inspired by something people often overlook — what happens after the trip. People capture amazing videos while traveling, but most of that content just sits unused. With WanderLens, those memories get a second life by automatically turning them into VR-ready videos that can help future travelers explore destinations more immersively.

As first-year students, this project was also a way for us to challenge ourselves and build something meaningful from scratch.

What We Built

WanderLens is an all-in-one travel platform built around five core features:

AI Trip Planner

A conversational planner where users describe their trip in plain language. It generates a structured, day-by-day itinerary including places to visit, timing, cultural insights, accessibility notes, budget estimates, and even photo tips. It also adapts based on user preferences like budget and travel style.

Real-Time Travel Search

Flights, hotels, and car rentals are integrated directly into the experience. Users can compare options, filter results, and move toward booking without leaving the itinerary view.

Interactive Map

Each stop in the itinerary is displayed on a live map. Users can explore nearby points of interest such as restaurants, museums, and viewpoints, along with useful details like names and opening hours.

VR Video Conversion

Users can upload travel videos, which are automatically converted into side-by-side VR format. These videos can then be viewed using VR headsets or simple cardboard viewers, making past trips more immersive.

Community & Gamification

Users can share reviews, photos, and VR videos. Contributions earn points, which can unlock travel discounts. A leaderboard highlights active contributors, and community content is integrated into trip planning.

How We Built It

Backend: Python with Flask, using SQLAlchemy and Flask-Login for authentication. The system is organized into modular components for clarity and scalability.
AI Integration: Google Gemini 2.5 Flash, with custom handling to ensure consistent and structured outputs.
Travel Data: Skyscanner Live Prices API for flights, hotels, and cars, with fallback data when needed.
Maps: OpenStreetMap with Nominatim for geocoding and Overpass API for points of interest, displayed using Leaflet.
VR Processing: FFmpeg for video conversion into stereoscopic format.
Frontend: Built with HTML and CSS, focusing on simplicity and performance without relying on heavy frameworks.

Challenges We Faced

JSON Reliability

AI responses were sometimes inconsistent. We built a custom parsing system with multiple fallback strategies to ensure stable, usable outputs.

API Limitations

The travel data API required a polling system rather than instant responses. We implemented a smooth fallback system so users always see results.

VR Conversion Quality

Creating a convincing VR effect required careful tuning of video processing parameters. After experimentation, we achieved a natural-looking result.

This project pushed us beyond what we had learned in our first year and gave us hands-on experience building a full, real-world application end to end.