JHANARICH Web Platform (Next.js)

Premium cookware-manufacturer website + admin console, built from "Website details.docx" and the official JHANARICH logo / WhatsApp product shots.

Run

cd ~/jhanarich-web
npm run dev        # http://localhost:3000

First-time setup (already done): npx prisma db push && npm run db:seed

• Website: http://localhost:3000
• Admin console: http://localhost:3000/admin → login admin / jhanarich2025 (dev defaults; change before deploying — override with ADMIN_USER, ADMIN_PASSWORD, ADMIN_SECRET env vars, see .env.example)

Stack

• Next.js 14 (App Router) + React 18, plain CSS design system (no UI kit)
• Prisma + SQLite (prisma/dev.db) — Product & Enquiry models
• Lenis smooth scrolling; all motion is dependency-light custom JS/CSS
• Fonts: Fraunces / Instrument Sans / IBM Plex Mono via next/font

Features

Website

• Awwwards-style editorial: Forged-for-fire hero (real pan cutout, logo medallion, steam, magnetic buttons), scroll-pinned triply exploded-layer scene with heat ripples and bond flash, kinetic marquee, glow-follow + 3D-tilt product cards
• Products render from the database; every card has an Enquire button that opens WhatsApp with a product-specific pre-filled message
• Enquiry form → saved to DB and opens WhatsApp with a pre-filled message
• Category filter tabs, counters, process timeline, OEM band, spec-sheet posters

Admin console (/admin)

• Overview: new/total enquiries, products live, catalogue composition bars, latest enquiries
• Enquiries: filter by status, status workflow (new → contacted → closed), WhatsApp one-click reply (when phone captured), delete
• Products: full CRUD — edit name/category/description/image inline, toggle live status & featured, reorder, add/delete products; site reflects changes instantly

AI layer (research-grounded: catalog-RAG with templated fallbacks, human-in-the-loop)

• JHANA assistant (floating chat, bottom-left): catalog-grounded Q&A — series tech, OEM, pricing process, address, shipping — with product cards attached to answers and one-tap WhatsApp handoff. Intent routing + keyword/retrieval matching runs fully offline; set GEMINI_API_KEY or OPENAI_API_KEY to upgrade phrasing quality automatically (same UX).
• Admin → Enquiries: ✦AI on any row classifies intent (oem/pricing/catalogue/horeca/partnership) + urgency, drafts a ready-to-send WhatsApp reply (editable, human-in-the-loop), or copies it if no phone. CSV export for the whole list.
• Admin → Products: ✦AI Write generates on-brand product descriptions (local template engine or LLM).
• Safety: outbound LLM calls are SSRF-hardened (https-only + host allowlist via lib/urlguard.js), chat is rate-limited (20/min/IP), enquiry input validated/clamped (lib/validate.js).
• Tests: npm test — 25 node:test cases covering validation, classification, chat routing, description generation, auth tokens and the URL guard.

Run

cd ~/jhanarich-web
npm run dev        # http://localhost:3000
npm test           # 25 tests

Optional AI upgrade: GEMINI_API_KEY=… npm run dev (or OPENAI_API_KEY=…). Without a key everything still works on the local engine.

Deploy

• Hostinger VPS (recommended — persistent SQLite, no external DB): see DEPLOY.md — one script sets up Node, nginx, PM2, DB and HTTPS.
• Serverless (Vercel): needs a hosted DB (Turso) — see commit ae2a55b notes; SQLite does not persist on serverless filesystems.

Legacy

The original static one-page build is archived at ~/jhanarich-site (served on :8940 while its python3 -m http.server runs).