The agent successfully assembled a comprehensive explanation of Cometly by combining direct site fetches with third-party research articles, since Cometly's own website relies heavily on JavaScript rendering and hides pricing behind a demo-booking gate. The agent retrieved substantive pricing and feature details from wetracked.io and usermaven.com, while also fetching Cometly's own blog post comparing attribution alternatives. The site was moderately navigable for discovery but deliberately obfuscated on pricing.

• ›The agent's direct fetches of cometly.com (steps [1], [3], [4], [13]) returned truncated HTML due to Next.js client-side rendering, making the site's actual content inaccessible without executing JavaScript. The agent correctly recognized this limitation and pivoted to third-party sources.
• ›Third-party pricing articles (steps [9] wetracked.io, [12] usermaven.com) were the primary content-bearing sources, providing concrete tier estimates ($44–$5,000+/month), ad-spend attribution, and feature breakdowns that Cometly's own site refuses to publish. These sources directly answered 'how it's priced' when Cometly's native pricing page does not.
• ›Step [13] (Cometly's own attribution-alternatives blog post) was cited and fetched, but also truncated, suggesting the site publishes positioning content but renders it client-side; the agent had to infer Cometly's differentiation from search snippets and third-party comparisons rather than Cometly's own articulation.
• ›The agent correctly called out that Cometly removed transparent pricing (as of April 2026 per third-party reports), meaning the site deliberately prevents agent discovery of a core commercial detail. This is a strategic choice that makes the site less agent-navigable for pricing-related tasks.
• ›Search results (steps [6], [7], [10], [15], [16]) were routing steps that pointed to fetchable resources but did not themselves contain the bulk of the answer; however, search snippets provided some feature context (e.g., 'server-side tracking', 'conversion sync', 'AI recommendations') that informed the agent's synthesis.