The agent successfully assembled a comprehensive explanation of Ditto by relying almost entirely on external sources and prior knowledge, since the ditto.ai website itself is JavaScript-heavy and did not render readable content in fetches. The agent gathered core information (what Ditto does, pricing, differentiation, target audience) from TechCrunch and other third-party articles, supplemented by search snippets and inferences. The site itself proved nearly non-navigable for automated retrieval.

• ›The ditto.ai domain fetches (steps [1], [3], [4], [5], [9]) all returned identical truncated HTML with no usable content — the site is a client-side rendered SPA that does not serve readable text to automated agents. The agent correctly identified this limitation and pivoted to search.
• ›All substantive information in the final answer came from web search results (steps [7], [10], [11], [12], [14]) and one successfully fetched third-party article (TechCrunch at [13]). The agent never extracted information from ditto.ai's own pages; it relied on secondary sources and prior knowledge (50% of fetches per metadata).
• ›The agent explicitly cited four sources in its final response, none of which are the primary site: TechCrunch, Ditto's homepage URL (mentioned as source but content not retrievable), StartupIntros funding page, and two blog reviews. The ditto.ai/pricing URL was fetched twice but returned no usable data, forcing the agent to infer 'completely free' from search snippets and articles.
• ›Key information gaps were identified by the agent (monetization timeline, role of human matchmakers, data privacy) — these represent content that even the third-party sources did not clarify, suggesting the site itself may not publish these details accessibly.
• ›The site's navigation structure (ditto.ai/pricing, ditto.ai/how-it-works, ditto.ai/about) was guessed or inferred, not discovered via a navigable homepage or sitemap; all six successfully fetched pages from the ditto.ai domain were routed via prior knowledge or direct URL guess rather than discovered links.