Safari Zbroia — tactical e-commerce on Django

Safari Zbroia is the online store of Safari Ltd in the tactical-gear segment. Roughly 3,000 SKUs are grouped into 1,084 categories and 208 brands. Two interface languages, full-featured filtering, cart, comparison, user accounts and order management — all running on a dedicated Django backend with PostgreSQL.

In ten years the project has gone through four engines — PrestaShop (2014), Joomla, OpenCart and Bubble (2024) — and finally settled on Django. All of its SEO equity, backlinks, indexation and organic traffic have been preserved through redirect maps assembled from the Wayback Machine archives and nginx access logs.

1C builds the catalogue its own way: every size is a separate SKU and the category tree is whatever 1C decides. The merchandising team needs a clean, web-friendly catalogue while prices and stock keep syncing from the accounting system. The task: combine automatic sync with full editorial control on the website.

The second part — rescue a decade of SEO after four engine migrations, deliver PageSpeed 95+, security hardening and readiness for AI search engines. All of it for a sensitive product segment with geographic traffic restrictions.

The business now has a single source of truth: 1C refreshes prices and stock every 15 minutes, while editorial overrides are protected by lock-fields. A single pass of the UK→EN auto-translator localised 2,119 products and 518 categories in ~40 minutes. PageSpeed climbed from 30 to 95+, HTML is 33% lighter and key images are served in WebP.

10,315 redirects preserved Google rankings after four engine swaps. The site is AI-search-ready: robots.txt allows OAI-SearchBot, Claude and Perplexity, /llms.txt describes the content map, and JSON-LD covers Product, ProductGroup, Brand, BreadcrumbList, OfferShippingDetails and MerchantReturnPolicy. Nonce-based CSP and a geo-block close the security loop.