E-commerce · On-device AI concierge
Luxora Commerce
A complete luxury e-commerce experience for watches and fine jewelry — a 33-page storefront, a full admin dashboard, and a rule-based AI shopping concierge that runs entirely in the browser. Zero runtime dependencies.
01 — The problem
Luxury retail online lives or dies on two things: a sense of craft in every pixel, and a browsing experience that helps people find the right piece without friction.
I wanted to build something that felt like a real, running boutique rather than a template — and to prove a genuinely useful AI feature could work with no backend, no API keys, and no framework at all.
02 — The process
01
The storefront
33 pages: faceted, URL-synced catalogue filtering, product pages with a zoomable gallery, a validated three-step checkout, wishlist, comparison and curated collections.
02
The AI concierge
A real rule-based engine that parses plain-language requests like 'a wedding gift under $4,000' into structured intent, scores the whole catalogue, and explains every recommendation.
03
The admin dashboard
KPI cards, revenue and region charts drawn as hand-authored inline SVG (no charting library), and sortable, searchable tables for products, orders, customers, reviews and coupons.
04
Zero-dependency architecture
Everything runs from a single window.LX namespace — no framework, no build step — working straight from the filesystem, backed by one generated data layer. Validated with a Playwright suite across all 33 pages.
03 — The impact
0
interactive pages, no framework
0
runtime dependencies
0
products, ranked by the AI engine
⚠︎ Placeholder metrics — replace with your real, verifiable figures before publishing.
Walkthrough
▶
Loom walkthrough — paste your embed link in content.py
