The agent successfully understood years.co by assembling information primarily from web search results and prior knowledge, since the site's own HTML pages failed to load readable content (Next.js client-side rendering issue). The agent retrieved pricing tiers, competitive positioning, and the clinic's unique research model from search snippets and cited blog posts, but had to work around the site's poor technical accessibility to answer the core task.

• ›Steps [1], [3], [11] (direct site fetches) returned truncated Next.js boilerplate HTML with no readable content—the site appears to rely entirely on client-side rendering, making it unreachable via server-side fetch. The agent was forced to rely on web search results instead.
• ›Steps [9], [13], [15], [19] (web searches) surfaced the site's own content via search indices and snippets: pricing pages, blog comparison posts ('YEARS vs. Neko Health', 'YEARS vs. AIVA'), and longevity.technology profile all appeared in search results and were cited in the final response. The site publishes this content but it is not directly accessible via HTML fetch.
• ›The agent correctly identified YEARS' three pricing tiers (€1,900, €7,600, €16,900), target audience, biomarker breakdowns, and key differentiators (clinic-as-study model, prospective clinical registry, physician-led diagnostics) despite never successfully loading the site itself—all information came from search-indexed content or prior knowledge.
• ›The agent candidly flagged seven gaps (exact biomarker lists, actionable outputs, visit duration, international accessibility, outcome metrics, research specifics, follow-up care) that were not available even in the discoverable content, demonstrating honest assessment of incompleteness.
• ›The site's technical architecture (Next.js client-side rendering) renders it agent-unfriendly for direct navigation but acceptable for discoverability via search engines, which indexed its blog and pricing content.