StyleGemma

Staring at a full closet with "nothing" to wear? Meet StyleGemma, your AI stylist that turns your "I don’t know" into "I look incredible."

Comment

Style Gemma - AI Personal Stylist

Your wardrobe, reimagined by AI

WHAT IT IS: Style Gemma is an AI-powered mobile app that transforms your existing wardrobe into endless outfit possibilities. Using Google's Gemini AI and Imagen 3, it analyzes your clothes and generates personalized outfit combinations with professional fashion photography.

KEY FEATURES:

  • Wardrobe Management - Upload and organize your clothes by category
  • AI Outfit Generation - Gemini creates 10-15 unique combinations per run
  • AI Fashion Photography - Imagen 3 generates professional fashion model photos
  • Smart Search - Filter outfits by color, style, or category
  • Favorites - Save your best looks
  • Body Setup - Personalized recommendations based on your measurements

TECH STACK:

  • Flutter (Dart) - Cross-platform mobile app
  • Google Gemini 2.0 Flash - Intelligent outfit generation
  • Vertex AI Imagen 3 - Fashion photography generation
  • Firebase (Firestore, Storage, Auth) - Backend infrastructure

HOW IT WORKS:

  1. User uploads wardrobe items with photos
  2. Sets style preferences (casual, formal, party, etc.) and body measurements
  3. AI analyzes wardrobe and generates 10-15 outfit combinations
  4. Imagen 3 creates professional fashion photos for each outfit
  5. User browses, favorites, and searches their personalized outfits

INSPIRATION: We built Style Gemma to solve the universal "I have nothing to wear" problem. Most people own plenty of clothes but lack creative vision to mix and match them. Our AI stylist makes professional styling accessible to everyone, eliminating decision fatigue and helping users rediscover their existing wardrobe.

CHALLENGES:

  • Integrated Vertex AI Imagen 3 with OAuth2 service account authentication
  • Managed API rate limits and quota exhaustion (free tier: 2 images/minute)
  • Solved Firestore 1MB field limit by uploading base64 images to Firebase Storage
  • Implemented smart fallbacks when AI quota exceeded

WHAT WE LEARNED:

  • Advanced Gemini AI prompt engineering for structured JSON output
  • Vertex AI authentication and API integration
  • Firebase real-time database architecture
  • Handling large binary data in mobile apps
  • Managing AI API quotas and rate limiting

FUTURE ENHANCEMENTS:

  • True virtual try-on with user's face/body
  • Weather-based outfit suggestions
  • Social sharing and outfit voting
  • Shopping recommendations for missing wardrobe pieces
  • Calendar integration for occasion-based styling

Built for Google AI Hackathon 2026 Tech: Flutter • Gemini AI • Imagen 3 • Firebase

Built With

Share this project:

Updates