- Task
- Build a premium site for a plastic surgery clinic: detailed service descriptions, a gallery of results, a blog and convenient online appointment booking, with an elegant design that conveys the level of the services.
- Solution
- The site is built as a React + Vite SPA with prerendering for search bots (as measured; the portfolio card says Next.js). Bots receive full HTML with a title, meta description, a single H1 and rich JSON-LD markup: MedicalClinic + Physician + BreadcrumbList + Occupation + WebSite/SearchAction. A correct sitemap is configured, along with Google Consent Mode v2 (default denied), an llms.txt acting as a content map for LLM agents, and a robots.txt with a split policy for AI crawlers: training crawlers (GPTBot, ClaudeBot, CCBot, Google-Extended, Bytespider, Meta-ExternalAgent, Applebot-Extended) are blocked, while search crawlers that cite the source (ChatGPT-User, OAI-SearchBot, PerplexityBot, DuckAssistBot, YouBot, MistralAI-User) are allowed.
- Result
- As of 31.07.2026 the site is live (HTTP 200, HTTPS): a search bot receives a fully prerendered page in 145 ms to the first byte and 178 ms to complete HTML (126 KB / 24 KB gzipped) with 10 JSON-LD types, including MedicalClinic, Physician, BreadcrumbList and WebSite/SearchAction — the fullest set of structured data among the 8 measured sites; the sitemap is correct, 89 URLs; there is also an llms.txt of 22,810 bytes and a robots.txt separating training from search AI crawlers — the only site in the sample with that.