The agent successfully assembled a comprehensive explanation of Merklebase by combining fragments from third-party sources, since the company's own website (.io domain) returned minimal extractable content. The agent learned that Merklebase is an institutional digital asset operations and portfolio analytics platform, identified its target users and key differentiators, but could not locate public pricing information. The site's heavy reliance on React with sparse server-rendered text made it poorly navigable for direct information extraction.

• ›Step [1] (www.merklebase.io/) fetched successfully but returned truncated HTML with no meaningful extracted text; the homepage itself was not content-bearing despite being the primary domain.
• ›Steps [10] and [11] (merklebase.ai/) returned the same non-extractable React boilerplate, indicating both .io and .ai subdomains use client-side rendering that obscures their actual value proposition from automated parsing.
• ›Steps [6] (LinkedIn company page) and [15] (CB Insights) became the de facto primary sources for concrete information about Merklebase's business model, competitors, and focus areas—not the company's own website.
• ›The agent cited sources [6] and [15] explicitly but could not cite [1], [10], or [11] as content-bearing, revealing a critical discrepancy: the primary domain is navigable (returns 200) but informationally hollow.
• ›The agent had to resort to 55% web search and 36% prior-artifact reliance (combining fragments across multiple third-party sites) to construct an answer, because the site's own pages lacked digestible, SEO-friendly, or machine-readable content.