FitMuse was inspired by the stress and decision fatigue many people feel when choosing what to wear, especially teens navigating anxiety, constant trend shifts, and an overflowing online fashion world. We wanted to bring back the fun and creative side of getting dressed by building an app that blends human imagination with AI rather than replacing it. FitMuse lets users upload their closet, choose a mood or vibe, sketch outfit ideas, or upload a character or celebrity image, and the app generates personalized outfits using real AI vision embeddings. Instead of hardcoding styles, FitMuse analyzes visual features from both the uploaded inspiration image and the user’s clothing, then finds the closest matches to recreate iconic looks directly from their own wardrobe. Building this required integrating image processing, similarity models, and a clean JavaScript interface while making sure the experience felt simple, intuitive, and creative. Our biggest challenges were extracting outfit details purely through AI, matching character looks to unique closets, and balancing automation with creative freedom. We’re proud that we built a functioning system that actually understands user clothing and generates real outfits rather than just mood boards. We learned a lot about embedding models, user-centered design, and how fashion and technology can meaningfully support self-expression. Next, we plan to add AR try-ons, a community styling space, smarter closet tracking, and partnerships that allow users to shop items that complement what they already own. FitMuse is just the start of a more creative, personalized, and accessible way to experience fashion.
https://github.com/vamsi678/fitmuse#- this is the link to the code
Log in or sign up for Devpost to join the conversation.