ora
capy.ai
success
Claude Code
intent

What does capy.ai do and who is it for? Explain it back to me.

11steps
48.5sduration
$0.1672cost
97,677tokens
16 steps5 reasoning steps5 searches
home
docs
docs
/features
docs
docs
search
search
search
search
search
100%
on-site discovery
67%
reliability
17%
link following
path origin
  • previous resource17%
  • prior knowledge83%
insight

The agent successfully understood Capy's core offering, pricing, and differentiation, but had to work around a largely client-side-rendered marketing site that exposed minimal content directly. Steps [1], [3], [7], and [8] returned real HTML, but the site's heavy reliance on JavaScript rendering meant the agent couldn't extract detailed information from those fetches alone—it relied heavily on prior knowledge (83%) and web search results to construct a coherent answer. The site is not agent-friendly: /docs and /features returned 404s, pricing and feature details required piecing together fragments, and critical information like credit unit costs and enterprise tiers remained opaque.

  • ›Steps [1] (homepage) and [3] (pricing) returned 200 responses but the HTML was truncated and heavily JavaScript-dependent, providing no actual readable content about what Capy does or how much it costs—the agent could not extract structured pricing tiers, feature lists, or differentiators directly from the fetched markup.
  • ›The agent discovered that /docs and /features paths do not exist (404 errors in steps [4] and [5]), forcing it to rely on web search and prior knowledge rather than canonical documentation or feature pages published by the site itself.
  • ›The agent's final answer draws heavily on web search results (steps [10–14]) and prior knowledge assertions about the three-tier agent architecture, parallel execution, and model flexibility—none of which appear to have been extracted from the actual HTML fetches, suggesting Capy's marketing pages do not publish this core product information in a machine-readable way.
  • ›The final response flags multiple gaps that confirm the site's opacity: credits unit costs are 'not 100% transparent,' enterprise pricing is 'mentioned but no specific pricing,' and 'no concrete data on how much faster development actually becomes'—indicating the site publishes positioning but not implementation details.
  • ›The agent's citation list includes direct links to /about, /pricing, /use-cases, and comparative articles, but the trajectory shows these pages were either guessed (not discovered via navigation) or discovered via web search results, not via site-internal discovery mechanisms.

Want to run your own?

Join the waitlist for early access to point your own agents at any domain, with the intents you choose.

or talk to us about agent readiness →