Login Page
Wardobe
Adding Clothes
Generating Outfits
Social Tab

Drip Management

Inspiration

Getting dressed should be easy, but most people either:

forget what clothes they own
struggle to match outfits
or waste time figuring out what looks good

We wanted to build an AI stylist that actually understands your wardrobe, not generic recommendations, but outfits made from your real clothes.

👕 What it does

Drip Management lets users:

Upload photos of their clothing
Automatically detect and categorize items using AI
Store a digital wardrobe
Generate full outfits based on their closet

The AI analyzes clothing attributes like:

type (jeans, hoodie, etc.)
color
material
style and fit

Then it creates cohesive outfits and even generates visual mockups of those outfits.

🛠 How we built it

We built a full end-to-end AI fashion pipeline using AWS + OpenAI:

Frontend & Auth

AWS Cognito for user authentication
React app hosted on AWS Amplify

Storage

S3 for storing clothing images and generated outfit images
DynamoDB for structured wardrobe + outfit data

AI Pipeline

User uploads clothing image
AWS Rekognition detects clothing labels
Lambda cleans and structures the data
Data stored in DynamoDB
OpenAI processes wardrobe data → generates outfits
OpenAI image generation creates outfit mockups
Images stored in S3 and displayed in the frontend

Backend

AWS Lambda for all processing
API Gateway / Amplify integration to connect frontend and backend

⚔️ Challenges we ran into

The hardest part wasn’t AI — it was wiring everything together.

Coordinating multiple AWS services (S3, Lambda, DynamoDB, Cognito, Amplify)
Managing async pipelines between image upload → processing → outfit generation
Debugging IAM permission issues
Bedrock InvokeModel access errors forced us to pivot mid-build

We originally planned to use AWS Bedrock, but due to IAM restrictions, we switched to OpenAI, which required us to rework parts of the pipeline quickly.

🏆 Accomplishments that we're proud of

Built a fully working AI wardrobe system end-to-end
Successfully integrated:
- image recognition
- structured data storage
- LLM-based outfit generation
- AI image generation
Created a system that turns real user clothing into styled outfits
Managed a complex multi-service AWS architecture under time pressure

📚 What we learned

IAM permissions can make or break your backend
Designing clean data flow between services is critical
AI outputs depend heavily on structured input
Rapid pivots (like switching from Bedrock → OpenAI) are part of real-world development