ChicMate is a multi-modal mobile application that provides personalized fashion outfit recommendations from user's wardrobe using state-of-the-art deep learning models. The app integrates image and text inputs—as well as location data—to offer context-aware style suggestions which suggests outfits that are apt for user's location in terms of weather and local trends. It features a conversational chatbot interface, digital wardrobe management, and real-time recommendations.
- Conversational Interface: Chat with the app to request outfit ideas.
- Image & Text Input: Upload images (e.g., a clothing item) or type queries.
- Digital Wardrobe: View, add, and delete wardrobe items.
- Context Awareness: Leverages past conversation history and geolocation to refine recommendations. (If users asks for recommendation for a specific location, it improvises accordingly).
- Personalized Recommendations: Uses CLIP for visual/text embedding and Google Gemini for natural language understanding.
- Frontend: React Native with Expo
- Backend: FastAPI
- Machine Learning: CLIP (via Hugging Face) for embeddings; Google Gemini for natural language generation
- Database & Authentication: Firebase Firestore and Firebase Authentication
- Location: Expo Location API
Navigate to the wardrobe screen to upload, view, and delete wardrobe items.
Use the recommendation chat interface to type queries or upload an image (e.g., a picture of pants) to get outfit suggestions.