Overview
ENVISION Wardrobe
iOS wardrobe app taken from concept to 90+ TestFlight users. Users photograph an item and Groq's Llama 4 Vision API auto-detects its type, colors, patterns, brand, and name — no manual entry. Self-hosted background removal on Google Cloud Run with rembg and Docker to cut per-image API costs entirely.
01
The Problem
People own dozens of clothing items but struggle to remember what they have, leading to redundant purchases and underused wardrobes. Manual wardrobe tracking apps require tedious data entry that kills adoption.
02
The Approach
I eliminated the data entry barrier by integrating Groq's Llama 4 Vision API — users snap a photo and the AI automatically detects the garment category, colors, patterns, and brand. The app organizes everything into a visual wardrobe with filtering and outfit planning.
03
Technical Highlights
- Groq Llama 4 Vision API integration that auto-detects an item's type, colors, patterns, brand, and name from a single photo
- Self-hosted background removal on Google Cloud Run with rembg and Docker, eliminating per-image API costs
- Firebase backend with Firestore for real-time wardrobe sync across devices
- Custom image processing pipeline to optimize photos before AI analysis
- TestFlight deployment pipeline with automated builds via Expo EAS
04
Results
90+ beta users on TestFlight
Concept to TestFlight in under 2 months
Positive user feedback on AI accuracy and ease of use
Preparing for App Store launch