StyleRise

Inspiration

The idea for StyleRise came from a common challenge many of us face—deciding what to wear for different occasions, whether it’s a job interview, a date, or even a casual outing. Hiring a traditional stylist can be both costly and time-consuming, making it inaccessible for most students and working professionals. Our goal was to create a simple, AI-driven solution that helps people make the most of the clothes they already own. Dress to Impress empowers users by allowing them to upload photos of their wardrobe and specify the occasion or style they’re dressing for. The app then uses AI to generate outfit suggestions, saving time and effort while boosting confidence without the need for a personal stylist or the purchase of new clothes.

What it does

Dress to Impress allows users to upload photos of their clothes, and the app prompts the user for the occasion of style they're dressing for. The app will then generate potential outfits.

How we built it

The frontend was developed using Next.js, providing a responsive and dynamic user interface. For data management, we utilized ChromaDB, a high-performance vector database, to efficiently store and retrieve clothing information. At the heart of the system, we integrated two AI agents connected via Fetch.ai. The first agent leverages the capabilities of Hyperbolic and Llama-3.2-90B-Vision-Instruct to analyze images of clothing from the user's wardrobe, extracting key attributes and generating detailed descriptions. These data points are then stored in ChromaDB, alongside user-provided information about the occasion and preferences to deliver personalized outfit recommendations of the day. The system also combines context relevant insights with other third-party APIs that to suggest additional items that would complement the user’s existing wardrobe. This tech-oriented cohesive architecture enables us to automate styling decisions, offering intelligent suggestions tailored to the user's needs and preferences.

Challenges we ran into

We faced several challenges throughout development, starting with communication difficulties within the team. An unexpected event involving theft created further disruption. Additionally, the team’s separation of tasks led to knowledge gaps, making it hard to align on certain aspects of the project. Lastly, we struggled to settle on a concrete idea before arriving at StyleRise. Despite these obstacles, we managed to overcome them by working together and staying focused on our shared goals.

Accomplishments that we're proud of

We are incredibly proud of each other and how we came together as a team. In addition to overcoming the challenges, we took the opportunity to learn new technologies and adapt quickly. Our ability to pivot and collaborate under difficult circumstances made this project particularly rewarding.

What we learned

Through this project, we gained valuable experience with technologies like Next.js, LLMs (Large Language Models), and modern databases. More importantly, we learned how to work effectively as a team and how to manage our time and energy throughout the event. We even figured out the importance of finding time to rest!

What's next for Dress To Impress

Looking forward, we aim to expand StyleRise by improving its functionality and user experience. Our next step is to build a user-friendly interface that allows users to easily manage and rate the outfits generated by the AI. A five-star rating system will provide valuable feedback that helps fine-tune the algorithm. Eventually, we plan to introduce a community feature where users can share, collaborate on, and remix outfit ideas. This will help us build a more engaging platform and foster a collective space for high-quality AI-driven fashion content.

Built With

• chromadb
• fetch.ai
• hyperbolic
• next.js
• ollama