Trivaro
Supplier-agnostic B2B flooring catalog — automated product feed engine, inquiry-to-quotation pipeline, Cloudflare-native stack
Trivaro is a B2B flooring supplier serving commercial, residential, marine, and aviation segments across Europe. The platform is a product catalog, not a store — buyers browse, collect products into an inquiry cart, and receive a formal quotation through a built-in CRM. It is designed to be supplier-agnostic: new product sources can be onboarded without changes to the frontend or the catalog structure.
Challenge
Flooring suppliers deliver product data in incompatible formats and update it on their own schedules. Any catalog that hardcodes one supplier's structure becomes a liability the moment a second one is added. The feed engine had to abstract that entirely — normalising, deduplicating, and indexing products from any source into one unified schema, automatically.
The B2B inquiry model doesn't fit standard e-commerce tooling. Buyers collect products over time, then request pricing — no checkout, no cart-to-payment flow. The platform needed its own inquiry pipeline: lead capture with product context, quotation and document generation, client management, and fulfillment tracking, all as a coherent internal tool built for this specific workflow.
Solution
The product engine runs as a Cloudflare Worker backed by serverless PostgreSQL and a hosted search index. Supplier feeds are ingested, normalised, and fingerprinted on a daily automated schedule — products stay current without manual intervention. The search layer supports faceted filtering by material, fire class, segment, and brand.
The inquiry pipeline converts browsing into a structured B2B workflow: the cart passes selected products into a lead with campaign attribution, which moves through client qualification, quotation, proforma, and invoice — with PDF generation, email delivery, VAT validation, and multi-warehouse fulfillment handled internally. The Astro 5 frontend runs in hybrid mode on Cloudflare Pages. Every category page carries structured schema data covering technical specs — fire classification, wear class, VOC, slip rating — so the catalog ranks for the queries B2B buyers actually search.