West Side AI

Case Study

Drape — AI Personal Stylist

Upload a photo, get 3 outfit recommendations with real brands and prices

3

Outfit recommendations per session

0 bytes

Image data stored

RAM-only

Privacy-first architecture

PROBLEM

The Challenge

Personal styling is either expensive (human stylist) or generic (quiz-based apps that ignore body type and skin tone). No tool could look at an actual photo, understand body proportions and coloring, and recommend real, shoppable outfits — until now.

SOLUTION

What We Built

ARIA (AI Recommendation & Image Analyst) — upload a photo, choose an occasion, and Claude Vision analyzes body type and skin tone to return 3 complete outfit recommendations with real brands, real prices, and direct shopping links. Privacy-first architecture: Multer with memoryStorage processes images in RAM only — nothing is ever written to disk or database. Sharp handles image preprocessing. Auth via Supabase (session logs only, no image data).

OUTCOME

The Result

Live at get-drape.com. Zero image storage architecture — no photos ever touch the database. Claude Vision delivers analysis in seconds. Supabase Auth with session tracking only. Deployed on Vercel (frontend) and Railway (Express API). The privacy-first design is a core product differentiator.

Tech Stack

Frontend
React 18, Vite 5, Tailwind CSS, Framer Motion
Backend
Express 4, Multer (memoryStorage), Sharp
AI
Claude Vision (Anthropic API)
Auth / DB
Supabase Auth (session logs only)
Deployment
Vercel (frontend) + Railway (API)
Privacy
RAM-only image processing, zero disk/DB writes

Want something like this for your business?

Work With Us