← All work

Case Study

Turning a warehouse of Magic cards into a one-click listing operation

A trading card resale team replaced hours of daily manual work with a single platform that identifies, prices, and lists cards across marketplaces automatically.

Industry

E-Commerce / Collectibles

Company Size

Small team (< 10)

Timeline

8 weeks

Tools Replaced

Manual process → 1 platform

Before

  • Every card identified by hand — name, set, edition, and condition checked manually against reference databases.
  • Pricing required cross-referencing multiple marketplace listings for each individual card.
  • Listings created one at a time on eBay and TCGPlayer with manual formatting and data entry.
  • A single team member could spend an entire day listing a few hundred cards.

After

  • Cards photographed and run through a computer vision system using OCR to identify name, set, edition, and condition automatically.
  • Low-confidence identifications routed to a human verification queue — a mechanical turk layer that keeps throughput high without sacrificing accuracy.
  • One click generates listings across eBay and TCGPlayer simultaneously — formatted, priced against real-time market data, and ready to publish.
  • Team throughput increased by orders of magnitude with the same headcount.

Hours

Saved daily per team member

2

Marketplace integrations

ML + Human

Hybrid verification pipeline

“We went from spending full days on listings to processing our entire inventory in a fraction of the time. The system paid for itself in the first week.”

— Operations Lead

Built with

React · Apollo/GraphQL · Node.js · MongoDB · Elasticsearch · Custom OCR pipeline

Have a similar problem?

Let’s talk about what purpose-built software could do for your team.