Infrastructure · 2024

Collage Tool APIAPI service.

A FastAPI service that ingests product photos, runs AI background removal, and assembles e-commerce-ready collages at scale. Stateless workers behind a queue, packaged in Docker.

Case study

Status
Live
Year
2024
Category
Infrastructure
01Challenge

What we were asked to solve.

E-commerce catalogues need consistent product imagery at volumes that don't justify per-image manual editing. SaaS providers in this space charge per call and bottleneck on rate limits when a brand wants to onboard a full catalogue in a weekend.

02Approach

How the studio took it on.

Python + FastAPI for the API surface, with image processing offloaded to worker containers running rembg and Pillow pipelines. Jobs go through a queue with idempotent processing so failed batches can resume without duplicating work. Dockerised so clients can self-host or call the hosted version.

03Outcome

What shipped, and what changed.

Serving multiple clients. Catalogues of several thousand products process in hours rather than the weeks they used to take by hand.