The problem
Sourcing or pricing parts across multiple brands and distributors means checking each one's catalog separately — different sites, different search interfaces, and for some distributors, a login wall in front of pricing and availability at all. That's the case for equipment and engine parts catalogs (Cub Cadet, Cummins, and parts brands like SMP, WDS and EXM) and, separately, for tech-distributor portals that only show real numbers once you're authenticated.
What was built
Two related builds. The first: a catalog scraper per distributor site, normalizing part numbers, descriptions and pricing into one searchable database instead of five separate ones. The second, a harder problem: a scraper against a session/cookie-authenticated distributor portal (Ingram/Techdata-style) — logging in, maintaining that session across requests, and extracting data that only exists behind the login, not on any public page.
Why this is the hard part
Public catalog pages are the easy case. Data behind a login is a different problem entirely: the session has to be established correctly, kept alive across a run, and re-established cleanly when it expires mid-way through — all without the distributor's site treating it as abuse. That's a genuinely different skill from straightforward scraping, and it's the same skill involved whenever a business's actual pricing or stock data only exists inside a supplier portal rather than on the open web.