Pasal

COVER
LOCAL MODEL PROCESS
EXAMPLE
EXAMPLE
CHECKOUT PAGE

Inspiration

This project was inspired by the common frustration of online shopping—buying clothes without knowing how they’ll actually look or fit. We also recognized how frequent returns lead to increased carbon emissions from packaging, shipping, and restocking. Pasal aims to solve both problems by making the try-on process digital, reducing guesswork and environmental impact.

HOW?

Pasal uses an advanced IDM-VTON model to seamlessly blend clothing product images with a personalized avatar, allowing users to virtually try on outfits and see how they would look in real life.

How we built it

We ran custom VTON models locally, testing different architectures and workflow to find the best balance between speed and realism. We integrated a model into a web store that aggregates clothing listings from multiple marketplaces, allowing users to virtually try on outfits directly at the checkout page.

Challenges we ran into

Running the VTON models on our own machines was tough—they're heavy and take a lot of fine-tuning to work smoothly. Another big challenge was creating APIs to pull live product listings from different marketplaces. Getting everything to work together in real time while keeping the experience fast and seamless wasn’t easy, but we made it happen.

Accomplishments that we're proud of

It works! We were able to work together as a team to create an end product we were happy with, and were also able to learn a ton about local models and how they work.

What we learned

We learned a lot about working with AI models in real-world environments, especially how to optimize them for limited hardware. We also got hands-on experience building APIs, handling messy data from different sources, and integrating everything into a smooth user experience. Most importantly, we saw how technology can directly improve sustainability in e-commerce.

What's next for Pasal

We plan to keep learning about and improving virtual try-on models to boost realism and speed. Our next goal is to explore how Pasal can evolve into a viable product by expanding marketplace support, improving fit accuracy, and making the experience seamless for everyday users.