Looksy

Inspiration

We've all been there - staring at our closet or scrolling through an online store, wondering “Will this outfit actually look good on me?” That mix of excitement and uncertainty can make shopping more stressful than fun. We wanted to fix that. Our idea started with a simple goal: help people feel more confident about what they wear - before they buy it. By combining realistic virtual try-ons with smart wardrobe tools, and powered by the FASHN AI’s Parallel-UNet tech, Looksy make outfit decisions way easier (and a lot more fun).

What it does

Virtual Try-On: Upload a photo and try on clothes virtually—yes, really. You can even try on multiple items in sequence to build a full outfit.
Virtual Closet: Keep track of your wardrobe by organizing items into categories like tops, bottoms, and dresses.
User Accounts: Secure logins and profile management so your closet stays your closet.
Image Management: Behind the scenes, we handle image uploads and retrieval efficiently, so everything stays smooth and snappy.

How we built it

Frontend:

React + TypeScript – For clean, type-safe code
Tailwind CSS – Because it makes everything look great, fast
Context API + React Router – Smooth state management and seamless navigation

Backend:

Express.js – Our trusty RESTful API server
MongoDB – Stores user data and wardrobe pieces
JWT + Bcrypt – Handles secure sessions and password protection

AI Magic:

FASHN API – With Parallel-UNet under the hood
Diffusion Models – For super realistic clothing renders
Cross-Attention + Implicit Garment Warping – To make clothes look like they actually belong on the person

Hosting & Deployment:

Amazon Web Services (AWS)

Challenges we ran into

A major challenge of this project is converting between formats and juggling various API requirements, as images come in all shapes and sizes, but the API can only handle some specific cases. Secure account storage also took a lot of research and trial-and-error. To avoid long processing times, we also spliced clothing images together so that it takes one generation instead of multiple for trying on multiple items.

Accomplishments that we're proud of

Building something that helps people feel more confident about their outfits - that’s the real win. But we’re also proud of the tech that made it happen. We’re especially proud of how the AI and engineering came together to create something both technically impressive and genuinely useful. The AI integration makes virtual try-ons look surprisingly realistic, while we tackled the image processing challenges to maintain quality and keep things efficient. On the user experience side, we focused on building a smooth, intuitive try-on flow with clear progress feedback, and a virtual closet that feels personal and easy to manage - plus, it all works seamlessly across devices. And behind the scenes, we made sure everything is secure and stable, with encrypted user data, protected images and passwords, and robust error handling to keep things running without a hitch.

What's next for Looksy

We’ve come a long way - but there’s still plenty we’re excited to improve. We're looking to speed up the multi-item outfit process, making it even smoother and more responsive. Social features are on the horizon too, like letting users share their favorite looks with friends or on social media. We also plan to introduce smart outfit recommendations based on personal style, along with analysis tools that can suggest tweaks or upgrades to complete a look.Integration with online stores is another key step, so users can shop directly from their virtual try-ons. On the experience side, we're aiming for better search filters, clearer clothing tags, and more detailed progress indicators during AI processing. Lastly, we’re working to make the try-on feature more inclusive by improving support for a wider range of body types.