London Spot website

London Spot sells British home and kitchen goods from Amazon liquidation lots through a shop and a warehouse in Tbilisi, almost all of them single units. The shop needed a site of its own that reads well on a phone in Georgian, is honest about each item’s condition and can be found on search engines, before payment and delivery are added.

Shoppers browse, filter and search, keep a wishlist and call the shop quoting the article code. The site is vanilla HTML, CSS and JavaScript with no dependencies; a Node build generates the English version under /en/, with the category tree and every product text, and runs i18n, link, syntax, width and SEO checks. It is hosted on an Oracle Cloud ARM server behind Caddy, with scripted deploys, checks and rollback.

A design pass set one layout system across five breakpoints, a condition label on every card, a call bar on phones and self-hosted Google Sans, verified with Playwright visual tests at 8 widths and with axe. The SEO build gives every product and category address its own pre-rendered head and first screen, plus sitemaps, JSON-LD, 4,820 product share cards, a product feed, GA4 through Tag Manager and Search Console. Both were run as orchestrated builds of AI coding agents, each task passing a fresh verifier.