FA©26

Case study — 2025

NDA · Shopify retailer

Supplier Catalog to Shopify, with AI

Role
Automation · AI
Service
Shopify
Status
Delivered to the client
In one line
Scrape a supplier catalog, enrich it with an LLM, upload to Shopify.

A Node/TypeScript pipeline that scrapes a wholesale supplier's product pages, uses an LLM to enrich descriptions, tags and categories, stores products in MongoDB and uploads them to a Shopify store.

01 — The problem

Listing a wholesale supplier's range by hand means copying a whole catalogue of product pages, rewriting every description so it isn't duplicate content, and tagging everything so shoppers can find it.

Client work under NDA. The system is described in full; the names stay private.

02 — What I built

  • Category scraper

    Crawls supplier category pages into product records.

  • LLM enrichment

    Rewrites descriptions and generates tags and categories.

  • Shopify upload

    Pushes products to the store through the Admin API.

  • Step-by-step CLI

    Separate scrape, process, upload and full-workflow scripts, with a progress API.

03 — How it runs

The flow — Crawl category pages, then Parse products, then LLM rewrites and tags, then Stored in MongoDB, then Uploaded to Shopify.
  1. Crawl category pages
  2. Parse products
  3. LLM rewrites and tags
  4. Stored in MongoDB
  5. Uploaded to Shopify

04 — Proof

  • Automated tests

05 — Stack

  • Node.js
  • TypeScript
  • Puppeteer
  • cheerio
  • OpenAI SDK
  • Shopify Admin API (REST + GraphQL)
  • MongoDB + Mongoose
  • Express
  • Winston
  • Jest

Next project

Yard Terminal

List once. Sell everywhere.