Lookas

Starting Page
Clothes search
Original photo before Virtual try on
Virtual try on
Outfit analyzer

Inspiration

Lookas was born from a common fashion frustration: you spot a stunning outfit on the street or social media, but have no idea where to get it. What if you could Lookas stylish, instantly? Upload the outfit, discover similar items from Inditex, virtually try them on, and get professional feedback on your style.

What it does

Lookas lets users:

Upload a photo of an outfit they love.
Use the Inditex Vision API to find similar clothing items from Inditex’s catalog.
Virtually try on those garments using advanced pose estimation and AI-based image fitting.
Get a professional outfit evaluation with feedback on style, fit, and overall impressiveness.
Buy the full look — quickly and seamlessly.

How we built it

We used:

Frontend: React + Next.js + TailwindCSS for a clean, responsive, and interactive UI.
Backend: Next.js API routes and server actions to handle requests and coordinate services.
Computer Vision: Integrated the Inditex Vision API for garment recognition and similarity matching, and developed a web scraping algorithm using Puppeteer to retrieve matching images from Zara's website.
Virtual Try-On: Implemented a pre-trained pose estimation model to generate image-to-image garment overlays.
Style Evaluation Engine: Integrated a pre-trained AI model to analyze multiple aspects of a photo, providing scores for fashion parameters like style, fit, and impressiveness.

Challenges we ran into

The Vision API doesn't provide image results, so we had to build a web scraper to extract visuals directly from Zara’s website.
Ensuring the searching and scraping processes were optimized and fast enough for a smooth user experience.
Training and tuning the virtual try-on model was challenging due to its complexity and our limited time.
Finding an effective and reliable tool to evaluate and provide meaningful outfit feedback was more difficult than expected.

Accomplishments that we're proud of

A fully functional virtual try-on prototype integrated with real outfit detection and image processing.
A polished, Zara-inspired brand identity with a sleek logo and modern UI.
Successfully integrating Inditex Tech’s Vision API into a real-time, user-friendly workflow.
Creating a product that bridges inspiration and action in one unified platform.

What we learned

How to effectively leverage fashion-specific computer vision APIs in a real-world application.
The importance of UX when combining AI-driven recommendations with user control.
How to build and coordinate a multi-layered pipeline that connects machine learning models, front-end interfaces, and backend services.
How to adapt quickly when APIs have limitations or when technical tools require workarounds.
The power of aligning brand aesthetics with functionality for maximum user engagement.

What's next for Lookas

Enhancing virtual try-on realism using 3D body models or by allowing users to input their own body measurements.
Adding a wishlist and sharing feature to save outfits or get feedback from friends.
Building partnerships with Inditex brands to enable direct checkout from search results.
Expanding outfit evaluation using generative AI (e.g., personalized suggestions: “Pair this with...” or “Try this instead”).
Launching a mobile-first version to let users snap and analyze outfits instantly, wherever they are.
Exploring sustainability insights, such as showing which similar items have lower environmental impact.