An SEO-first OEM parts catalog
Exploded diagrams with clickable callouts, fitment data and quote requests — pages served in milliseconds
The problem
Buyers of agricultural machinery parts need the exact OEM part number for their model and assembly, the way a dealer catalog shows it. The owner needs that traffic to come from search engines and AI assistants and turn into quote requests.
What we built
- Browse brand → model → assembly → diagram → part, with clickable diagram callouts
- Part pages with factory data, model fitment, supersessions and alternatives — sourced strictly from OEM data
- Part-number search and bulk quote parsing: found / superseded / similar / needs clarification
- Markdown articles that link straight into the catalog, with Article, FAQPage and BreadcrumbList schema
- SEO and GEO: 45,000-URL sitemaps, IndexNow, llms.txt, analytics goals, generated OG images
- Bot policy: search engines and AI assistants verified by reverse DNS and vendor IP lists; AI training crawlers and SEO scrapers blocked
Engineering
Cold part page: 1.2 s → 8 ms
- Three-layer caching (versioned on-disk HTML, cached heavy queries, background warm-up): a cold part page dropped from 1.2 s to 8 ms, and a 15.8 s query is served from disk
- Built on a catalog database with a 12.4M-row unindexed table — covering indexes and split queries keep every page within its time budget
- Node.js 22
- Framework-free SSR
- PostgreSQL
- Vanilla JS
- CSS design system
- JSON-LD
- IndexNow
- LLM API