PittDrip
Inspiration
The inspiration for PittDrip stems from a common dilemma: standing in front of a packed closet yet feeling like you have nothing to wear. In our fast-paced lives, deciding what to wear can become a stressful task, especially when considering the weather and planned activities. We wanted to create a solution that not only simplifies this process but also adds a layer of excitement and creativity to outfit selection. By harnessing the power of AI, PittDrip aims to be your ultimate closet buddy, transforming how you interact with your wardrobe.
What it does
PittDrip is an innovative AI-driven closet assistant designed to help users curate and manage their clothing collections effortlessly. The app allows you to upload images of your garments, which are then automatically categorized into types, colors, and styles. But it doesn’t stop there! PittDrip intelligently generates personalized outfit suggestions based on real-time weather conditions, specific activities, and your unique closet inventory. Whether you’re dressing for a casual outing, a formal event, or a workout, PittDrip ensures you look your best while staying comfortable and stylish.
How we built it
We built PittDrip using a combination of cutting-edge technologies to deliver a smooth and engaging user experience. The front-end is developed with React Expo, which enables us to create a mobile-friendly application with rich interactive features. For the intelligent outfit recommendations, we utilized Gemini multimodal LLMs that analyze both visual and contextual data.
Our custom models are deployed via Replicate, which facilitates seamless interactions between various AI components. For user authentication and data storage, we chose Firebase/Firestore, ensuring that users' information is secure and easily accessible. To provide timely weather updates, we integrated the OpenWeatherMap API, allowing us to factor in weather conditions when generating outfit suggestions. Our default recommendation model, Roc the Panther, serves as the backbone of our outfit generation process, leveraging advanced machine learning techniques.
Challenges we ran into
The journey to develop PittDrip was not without its hurdles. Some of the key challenges we encountered included:
CORS Issues: We faced significant Cross-Origin Resource Sharing (CORS) problems that hindered our ability to make API calls effectively. Debugging these issues required extensive research and testing.
Image Uploads: Sending images over HTTP proved to be more complex than anticipated. We encountered difficulties in ensuring that images were processed and categorized correctly, impacting user experience.
Chaining Models and APIs: Integrating multiple models and APIs to create a cohesive and efficient workflow was a daunting task. Ensuring smooth transitions between these components required careful planning and troubleshooting.
Devpost Mishap: A particularly frustrating incident occurred when our entire README was deleted on Devpost, forcing us to reconstruct our documentation from scratch under a tight deadline.
API Key Leaks: We learned the hard way about the importance of securing API keys, as several were inadvertently exposed during development. This experience underscored the necessity of implementing robust security measures.
Accomplishments that we're proud of
Despite the challenges, we are proud of what we achieved with PittDrip. We successfully developed a functional prototype that accurately categorizes clothing items and generates stylish outfit recommendations tailored to user needs. Our ability to integrate various technologies—AI models, APIs, and cloud services—into a cohesive application speaks to the dedication and collaboration of our team. Additionally, we have received positive feedback from initial users, which motivates us to further refine and enhance the app.
What we learned
Our journey with PittDrip has been a profound learning experience. We gained invaluable insights into the intricacies of web development, particularly regarding secure coding practices and the importance of safeguarding API keys. The challenges of CORS and HTTP image uploads taught us to think creatively and adopt problem-solving approaches. Moreover, we learned the significance of teamwork and communication, as we navigated setbacks and celebrated our victories together. Each challenge reinforced our commitment to continuous learning and improvement.
What's next for PittDrip
Looking to the future, we have an ambitious roadmap for PittDrip that includes:
Enhanced Machine Learning Models: We plan to develop more sophisticated algorithms that take into account user preferences, trends, and style profiles to offer even more personalized outfit suggestions.
User Profiles and Preferences: By incorporating user profiles, we can tailor recommendations based on individual tastes and past outfit choices, creating a more engaging experience.
Improved Security Measures: Strengthening our security protocols to protect user data and API keys will be a top priority, ensuring that user trust remains intact.
Social Sharing Features: We aim to enable users to share their favorite outfits on social media platforms, fostering a community of fashion enthusiasts.
E-commerce Integration: Exploring partnerships with online retailers to provide shopping recommendations based on user wardrobes, making it easier to find new items that complement their existing collection.
Join us on this exciting journey to redefine the way you approach your wardrobe. With PittDrip, we’re not just helping you choose outfits; we’re empowering you to express your unique style with confidence!
Log in or sign up for Devpost to join the conversation.