B2B Platform

shwcs. From a scattered vendor catalog to an editorial platform where every solution is explained and connected to whoever runs it.

We built shwcs, a B2B discovery platform: solution publishing with editorial review, a personal library to save and compare, consent-based contact requests, and replies from the owner account. Next.js 15, Neon PostgreSQL, and semantic search with pgvector.

2 Locales (ES/EN)
3 Listing types
100% Editorially reviewed
pgvector Semantic search
shwcs — Flouvia
THE CHALLENGE

Mexican B2B software, agencies and services live scattered across directories with no criteria: listings without context, no idea what each option solves or how to reach whoever runs it. There was no place where every solution is explained and contact happens with consent, not through email scraping.

THE SOLUTION
  • Next.js 15 (App Router) platform with locale routing and mirrored ES/EN content.
  • Solution publishing with editorial review before entering the catalog.
  • Personal library: save, compare and organize options.
  • Contact requests with explicit consent; the owner replies from their own account.
  • Semantic search with pgvector: relates solutions by what they solve, not just by keyword.
  • Persistence in Neon PostgreSQL with Drizzle ORM; files in private Vercel Blob.
TECH STACK
Next.jsPostgreSQL (Neon)Drizzle ORMpgvectorVercel BlobAI SDK
THE IMPACT

Platform in operation with an initial catalog (includes Cord and Flouvia as owned listings). The figures describe the delivered architecture and scope; traffic and vendor-activation metrics will be added as the catalog grows.

2 Mirrored content locales

Spanish and English served from locale routing, without duplicating templates: the catalog, listings and contact flows stay aligned in both languages.

Opt-in Consent-based contact

A prospect requests contact explicitly and the owner replies from their account. No exposed emails, no scraping: the data is shared only when both sides agree.

pgvector Semantic discovery

Embeddings live in Postgres (pgvector), so relating solutions by what they solve needs no separate search service and no external index to sync.