StyleVerse

Note: At the time of video submission, we were unable to capture the full capability of our app as we ran into an unforeseen error that has since been fixed.

Inspiration

Overconsumption and fast fashion are huge issues in today's world. If someone is going on vacation and looking for summer looks for their trip to Europe or if they need to go to a Halloween party, they be tempted to buy everything they need and forget what they can reuse within their own closet. We hope to encourage users to instead reuse outfits already in their closets that still fit their desired style and use cases.

What it does

The user can upload clothes from within their own closets and easily view everything that is in their closet. They can also views the clothes that their friends have through a shared closet. Then, the user can specify context for their outfit like the weather (e.g. cloudy), the size of their clothes (e.g. medium), and the style they are looking for (e.g. cozy fall bookworm vibes) to get an outfit that is completely personalized to their unique needs and incorporates clothes from their closets and their friends' closets. The user can see what they may look like wearing the outfit in an AI-generated image as well.

How we built it

We used MongoDB to store the users and the images in each of their closets. We coded mainly using Python with VSCode and we used Streamlit to create and deploy an interactive site. We used Gemini 2.5 Flash Image (Nano Banana) to power image creation, which is accessible through the Gemini API.

Challenges we ran into

We faced many challenges throughout including setting up the MongoDB database and figuring out how to implement a network of people within it and issues trying to get Nano Banana to generate images the way we expect it to, using only the clothes available in the user's closet.

Accomplishments that we're proud of

We are especially proud that since this is most of team's first hackathon, that we were able to build a product that works to some extent and is one we are able to demo. We are also proud of implementing MongoDB, because we were not sure if we could figure it out.

What we learned

We learned how to work with different technologies, how to collaborate with a team, and how to persevere past obstacles that make you want to just give up.

What's next for StyleVerse

We love this idea so much because there are so many directions we can take this in the future. We really want to have a robust network of friends that the user has that the model can draw upon because one of our main focuses is encouraging looking at the clothes you and the people around you already have. We want to also really optimize the image generation so it is accurate to the user and useful to visualizing how the outfit would look on them.