Skip to content
House of Hazi
Back to projects
Shipped 2026-07-24

Sosnowscy Wypieki Domowe — Bilingual Patisserie Website & Custom CMS

A bilingual (Polish/English) website and lightweight custom CMS for a family-run patisserie in Poland, built with Next.js and engineered to be discovered by both traditional search engines and AI answer engines like ChatGPT, Perplexity, and Google AI Overviews.

Stack
  • Next.js
  • TypeScript
  • Tailwind CSS
  • next-intl (i18n)
  • Vercel (Hosting, Blob Storage, Analytics)
Category
Cafe & Bakery
Live
Visit site
Click to enlarge
1 / 6
Sosnowscy Wypieki Domowe is a two-generation family patisserie (cukiernia) in Nowy Bedoń, near Łódź, Poland. The goal was to give them a professional, fully bilingual web presence, backed by a simple content system the owners could manage themselves without touching code, while positioning the site to actually be found: by customers searching Google, and increasingly by AI-driven search and recommendation tools. Bilingual by design. Every page, from the homepage to individual product listings, is fully localized in Polish and English using next-intl, with locale-aware routing (/pl/..., /en/...), hreflang alternates, and per-page metadata so search engines correctly serve the right language to the right audience. A CMS without the CMS overhead. Rather than standing up a heavyweight headless CMS for a small single-owner business, gallery items (cakes/pastries), news posts, and the homepage hero image are managed through a custom-built admin panel, authenticated with JWT sessions and backed by Vercel Blob storage. It's intentionally lightweight: fast to load, easy to reason about, and cheap to run. Built to be recommended, not just indexed. This is the part I'm most proud of: beyond standard SEO (dynamic sitemap, robots.txt, Open Graph tags), the site embeds schema.org structured data (LocalBusiness, Product, BreadcrumbList) on every relevant page, so search engines and AI systems can read exact, structured facts, hours, address, prices, product details, rather than having to infer them from prose. It also ships an llms.txt file, an emerging convention that gives AI crawlers a direct, unambiguous summary of the business, and its robots.txt explicitly welcomes AI crawlers (GPTBot, ClaudeBot, PerplexityBot, Google-Extended, and others) rather than leaving that decision ambiguous. Compliant by default. A bilingual privacy policy was written to match exactly what the site actually does, no boilerplate copy-paste, covering hosting logs, the language-preference cookie, and the Google Maps embed, appropriate for a GDPR-covered EU business with a minimal data footprint. Production-hardened. Vercel Analytics and Speed Insights are wired in for real usage data, and a subtle-but-real hydration bug (duplicate <html>/<body> declarations across nested layouts, a common Next.js App Router pitfall) was diagnosed and fixed by testing an actual production build locally before redeploying, catching an issue that only manifested in production, not in development. The result is a fast, fully bilingual, professionally structured small-business site that's realistic for its scope, and quietly ambitious in the parts that matter for long-term discoverability.