Inspiration

Our inspiration was Clueless and the outfit generator that was featured. We also wanted to keep overconsumption in mind, so we created this app to help users become re-inspired by their wardrobe before shopping. Furthermore, we have a sustainability mode so that if a user doesn't have specific item, our style assistant can direct users to products that are sourced from second hand websites.

What it does

StyleMe is an AI fashion assistant that can help you build outfits! StyleMe works in two primary modes: wardrobe and shop. The wardrobe mode allows a user to input images from their closet alongside a style intention and StyleMe will build an outfit according to a user’s items and style preference. Shopping mode allows a user to input a style intention and then search the web for an entire outfit. Within shopping mode there is a sustainability mode in which the user will be returned items from second-hand websites. These various functions allow a user to interact with our AI style assistant to build an outfit.

How we built it

Wardrobe Mode: We use Gemini to produce an outfit suggestion based on the images inputted by the user and the provided user request. We then display Gemini’s text recommendation of the outfit alongside the corresponding subset of images from those inputted by the user. Despite Gemini not being able to output images yet, we by-pass this by figuring out which images the model has recommended using our AI-built dictionary and displaying these, along with the outfit description! Shopping Mode: We use Google Search APIs to retrieve product links for the customer based on the inputted style. For sustainability mode, we built a database of clothing items that map to sustainable shopping links, due to the current unavailability of these websites’ APIs. Graphical User Interface: We built this using React, Javascript, and connected to our backend using Flask.

Challenges we ran into

We were unable to get access to APIs by certain websites so we had to create an artificial database of clothing items and links to sustainable products. We really wanted to be able to output a visual of the completed outfit, which was challenging because Gemini doesn’t output images yet. We overcame this by creating an AI-labeled dictionary in the pre-processing stage which we later used to convert Gemini’s text results (outfit recommendation) into image output.

Accomplishments that we're proud of

We're proud of successfully creating a variety of different modes for users to both browse their wardrobes and shop online, as well as specifically creating a sustainability mode that encourages users to shop green as this aligns with our values. We also created multiples for the user to interact with our feature, including text input, image input, and url input! We're proud that we were successfully able to leverage Gemini to make a useful and socially impactful service that we're passionate about in just 2 days!

What we learned

We learned how to interact with Google's Gemini APIs and how we can output this to a graphical user interface so that users can interact with Gemini. We also learned how to build a full-stack application from scratch.

What's next for StyleMe

StyleMe can be integrated with more APIs from various websites so that we can search for items on various websites directly. StyleMe also wants to expand to accepting video format, so that you can upload your favorite style influencer's videos and get an outfit that they would wear. We could also integrate this product for businesses so that users can have an outfit based off of the items from a website and a style intention.

Built With

Share this project:

Updates