SAATATYA LABS
Enquiry

Case study · DRS Blouse

Wholesale and retail, one storefront.

DRS sells designer blouses to wholesalers and retail buyers from the same catalog. The storefront prices every product three ways, feeds Meta clean conversion data, and runs its whole catalog operation from one admin, live at drsblouse.com.

Running · drsblouse.comFastAPIPostgreSQLReact + ViteCelery + RedisImageKit

113

API endpoints

21

Database tables

3

Price tiers

12

Repositories

Station 01 · The buyers

Two kinds of buyer, one catalog.

Wholesale buyers with GST numbers, retail shoppers and guests all browse the same designs at different prices. Every product carries a wholesale, retail and guest price, and orders wait for admin approval before stock moves, so the factory confirms what it can actually ship.

orders · approval flow
BROWSECARTORDERHOLD← admin approves
Role-based pricing

Station 02 · The catalog

A variant matrix that runs itself.

Every design exists in a matrix of colors and sizes. The admin uploads in bulk, images process in the background with live progress, failures park in a dead letter queue instead of vanishing, and ImageKit serves every photo at the right size.

catalog · bulk upload
$ upload catalog
→ images queued to Celery
→ color and size variants built
✓ CDN URLs stored
✓ failures parked in DLQ
Bulk upload, zero loss

Station 03 · The signal

Marketing data Meta can trust.

The browser pixel and the server Conversions API fire the same events with a shared id, so Meta deduplicates instead of double counting. Purchases are fired only by the server, never the browser, and personal data is hashed before it leaves.

signal · conversions
BROWSER EVENTS 6PURCHASE SERVERPII SHA-256
Server-authoritative

Station 04 · The architecture

Small pieces, cleanly separated.

A React storefront on Vercel talks to a FastAPI core on Render through a repository layer, with Redis for cache and pub sub, Celery for image work, and WebSocket updates streaming into the admin as things change.

system map · storefront
REACT SPAFASTAPIREDIS + CELERYPOSTGRESQL
Render + Vercel + Redis

Station 05 · Still running

Live at drsblouse.com.

The storefront runs in production with security headers and health checks across database, cache and CDN, while the same client's factory floor runs Inventory OS next door. One client, two systems, one team.

status · live
ENDPOINTS 113TABLES 21STATUS RUNNING
● Running now

System map · verified

The architecture, on paper.

Shopfront

React SPA

storefront + admin

Core

FastAPI on Render

23 routers · repository layer

Celery worker

image processing only

Data

PostgreSQL

21 tables

Redis

cache + broker

Outside

ImageKit CDN

originals + variants

Meta CAPI

server-side events

Arrived. Now let's chart yours.

Every client gets a new flight path. Bring us the destination and we will engineer the route.