An eBay Reselling Engine for Auto Parts

The Challenge

An auto-parts reseller had a 386,561-product wholesale catalog but no way to know which items were actually profitable to list on eBay. Pricing against competitors and creating listings was all manual — so most of the catalog never got sold.

The Approach

We built a SaaS platform that owns the whole loop: bulk-import the vendor catalog into an indexed database, scrape competitor eBay stores with a browser bot, match the same part four different ways, calculate profit, and push priced listings to eBay through its API — with repricing rules that keep a minimum margin as competitor prices move.

How we built it

1. 01

   Ingested the catalog

   Bulk-imported a 386,561-product vendor feed (TXT / CSV / Excel) into an indexed SQLite schema of 42 tables, tuned with WAL mode and caching for speed.

2. 02

   Scraped the competition

   A Playwright browser bot scrapes competitor eBay stores — with anti-detection and captcha handling — to pull live prices and fitment.

3. 03

   Matched products four ways

   Tiered matching (partslink, OEM, part-number, then fuzzy keyword) reliably finds the same part across two different catalogs.

4. 04

   Priced for profit

   A repricing engine with four rule types (margin %, fixed markup, undercut-lowest, match-average) prices every item above cost with a guaranteed minimum margin.

5. 05

   Listed & tracked on eBay

   Approved items push to eBay via the Inventory API over OAuth2; orders sync back and profit is tracked live in the dashboard over WebSockets.

What We Built

• 18 FastAPI routers, 42-table SQLite database (WAL mode)
• Playwright scraper with anti-detection & captcha handling
• 4-tier product matching across catalogs
• 4 repricing rule types with minimum-margin guarantee
• Gap / hot-product / dead-stock intelligence reports
• Real-time listing queue over WebSockets

Connected Platforms

From the project