Danny Cen

Nova Imports — B2B/B2C ecommerce with AI

Headless commerce for imported goods: catalog, cart, wholesale rules, and AI-assisted product intake — PostgreSQL, Redis, and Docker.

  • Role Full-stack architecture & implementation
  • Studio Danny Cen
  • Period 2025–2026
Nova Imports — B2B/B2C ecommerce with AI

The project

A B2B/B2C ecommerce platform for Nova Imports (wholesale and retail imported products). Headless architecture: MedusaJS v2 for commerce APIs and Next.js for the storefront.

Includes a wholesale module (pieces per box, MOQ, sell type), tiered price lists, and an AI product extraction endpoint (Gemini Vision) to speed up admin catalog intake.

The problem

Running wholesale and retail in one store needs MOQ and tier pricing that a generic theme does not cover well. Catalog intake from photos/specs was also slow.

Architecture (high level)

  1. MedusaJS v2 — cart, checkout, admin, wholesale-info module, MOQ middleware.
  2. Next.js 16 (App Router) + React 19 + Tailwind — catalog, PDP, account, checkout.
  3. PostgreSQL 16 + Redis.
  4. Optional Stripe; Gemini for POST /admin/ai/extract-product (mock without API key).
  5. Docker Compose for staging/production.

My contribution

  • Headless architecture and custom wholesale modules.
  • MOQ validation in the API and admin widgets.
  • Next storefront and checkout flow.
  • AI product intake and demo seeds.
  • Docker packaging and ops docs.

Visible outcomes

  • Store operable with B2B rules (MOQ / boxes / tiers).
  • AI-assisted product intake (or mock locally).
  • Stack ready for custom domains (store + API).

Stack (examples, not a ceiling)

MedusaJS v2 · Next.js 16 · React 19 · Tailwind · PostgreSQL · Redis · Stripe · Gemini Vision · Docker.

  • MedusaJS v2
  • Next.js
  • PostgreSQL
  • Redis
  • Stripe
  • Gemini Vision
  • B2B / MOQ
  • Docker