Your closet-aware fashion critic.
FitCheckAI is a personal AI stylist that tags your clothes, roasts your fits, and builds outfits - all through an interactive Streamlit UI.
🎥 Demo Video:https://youtu.be/s57yTOkLLwY
Upload a photo of your outfit and get a brutally honest AI critique, including:
- Style breakdown
- Score out of 100
- Savage one-liner
All results are saved to MongoDB and de-duplicated using perceptual hashing.
Upload individual clothing items - the AI classifies:
- Item type (
shirt,shoes, etc.) - Color (
black,white, etc.) - Setting (
indoor/outdoor) - Formality
- Gender style
Saved in a consistent JSON format and stored in /Closet.
Generates full outfit recommendations based on:
- Location (indoor/outdoor)
- Formality (casual/formal)
- Gender
- Preferred color
Color compatibility rules are applied (e.g. blue pairs with beige, white, brown, etc).
| Component | Tool / Library |
|---|---|
| UI | Streamlit |
| Fashion Critique | LangChain + Qwen2.5 VL (3B, quantized) |
| Image Tagging | CLIP (openai/clip-vit-base-patch32) |
| Database | MongoDB Atlas |
| Data Format | JSON / .jsonl |
.
├── fitcheck/
│ ├── Fashion AI Advisor.py # Main Streamlit app
│ ├── analyze_outfit.py # Outfit critic logic
│ ├── tagging.py # Tagging with CLIP
│ ├── tag_closet_items.py # Bulk closet tagger
│ ├── test_Analyze_Outfit.py # Unit test for LangChain outfit critique
│ ├── test_tag.py # Unit test for tagging module
│ ├── testmongoconnection.py # MongoDB connection test
│ └── pages/
│ ├── 1_Add_to_Inventory.py # Upload and tag closet items
│ └── 2_Get_Outfit_Suggestion.py# Outfit recommender (rule-based)
├── Closet/ # JSON-tagged clothing items
├── Images/ # Outfit photos
├── designs/ # Wireframes + mockups
├── requirements.txt
├── run_app.bat
└── vlm_tagging_test.ipynb # Prototype testing
Style: The outfit features a casual yet trendy ensemble consisting of a black beanie hat, a white scarf wrapped around the neck, a striped long-sleeve shirt under a black vest over jeans. The combination suggests an urban, laid-back vibe but lacks depth due to its simplicity.
Rating: 35/100
Comment: "This outfit might as well have been designed by a robot; it's so formulaic."
{
"image_id": "abc123",
"item_type": "Jacket",
"color": "Black",
"indoor_outdoor": "Outdoor",
"formality": "Casual",
"gender": "Men's",
"path": "Closet/jacket.jpg",
"folder": "Closet"
}
These were mockups used to guide layout and flow.
Landing Page
Upload Interface
LeCritique Display
Install dependencies:
pip install -r requirements.txt
Then launch the app:
streamlit run "fitcheck/Fashion AI Advisor.py"
- Daniel Mehta
- George Fotabong
- Dylan Higuchi
- Jacob DeAbreu
- MongoDB URI is hardcoded for now (demo only - no sensitive data).
This code is free to use, modify, and share for non-commercial purposes.
Commercial use requires prior written permission. All rights reserved.