The agent successfully assembled a comprehensive overview of vaults.fyi by combining fragments from documentation, a blog post, and inferred comparisons from web search results. The site's information was distributed across multiple resources (docs subdomain, blog, API page) rather than centralized on the homepage, forcing the agent to conduct multiple fetches and web searches to answer what vaults.fyi does, who it's for, how it's priced, and how it differs from alternatives. Pricing details were found but incomplete; competitor comparisons were largely inferred rather than explicitly published by the site.

• ›Step [8] (docs.fyi/introduction) and Step [14] (vaults.fyi/api) provided core functionality and use-case information, but were truncated HTML responses that the agent had to interpret. No structured data (JSON-LD, microdata) was visible in the returned HTML.
• ›Step [13] (blog.vaults.fyi blog post) was a direct hit—the only content-bearing step that explicitly explained 'What is the vaults.fyi API?' in narrative form. This suggests the site relies on blog content for product explanation rather than homepage or dedicated landing pages.
• ›Step [15] (methodology page, fetched via web search) delivered the key differentiator insight: 'onchain data, not scraped.' This critical positioning claim had to be discovered via search, not surfaced prominently by the main site.
• ›Pricing information (Step [6], docs.fyi/plans/overview) was available but the agent noted it was vague—tier names and prices were listed but credit allocation per tier and rate limits were not explained, making cost planning unclear.
• ›Step [16] (customer case studies) and Alchemy alternatives list provided indirect competitor and customer signals, but no published feature-by-feature comparison table exists on vaults.fyi itself. The agent had to infer differentiation from fragments.
• ›The homepage (Step [0]) and main API page (Step [14]) both returned truncated HTML, preventing the agent from assessing whether full positioning and pricing were visible there or if the site deliberately depends on linked resources.
• ›Steps [3] and [4] (pricing and blog route attempts) returned 404 errors, suggesting the site may have restructured or the agent's URL guesses were incorrect—a sign that navigation paths are not intuitive.