The agent successfully assembled a comprehensive explanation of Meilisearch by relying heavily on web search results and the company's own blog comparisons, rather than the main website pages. The site's JavaScript-heavy frontend proved unreadable to the agent, forcing it to pivot to search engines and third-party sources to answer the task. Despite this limitation, the agent delivered a complete answer covering what Meilisearch does, its pricing tiers, target audience, and differentiation from competitors—though some details like exact pricing numbers and enterprise plan specifics had to be pieced together from fragments.

• ›The main site pages (homepage, pricing, features, docs) all returned HTML that was too JavaScript-heavy for the agent to parse meaningful content, causing it to pivot to web search after [5].
• ›Web searches [6] and [7] directly returned Meilisearch's own blog URLs in their snippets, which proved far more useful than the official pages. The agent then fetched [8], one of those blog articles (algolia-vs-elasticsearch), which was explicitly cited in the final response and appears to have been content-bearing, though the HTML was similarly truncated in the response.
• ›The bulk of the substantive answer (pricing plans, positioning vs. Elasticsearch/Algolia, use-case fit) came from search results and prior knowledge, not from the fetched pages themselves. The agent noted that pricing details 'weren't clearly spelled out on the pages I could access' and had to source specifics from third-party sites like aiproductivity.ai and elestio.
• ›The agent correctly identified gaps in what it could find: unclear cloud hosting details, vague 'AI Retrieval' positioning, missing enterprise pricing, and obscured product roadmap. These represent real absences from the site itself, not just agent blindness.
• ›The site's agent-readiness was poor for automated extraction: the homepage, pricing page, docs, and blog all returned JavaScript-rendered HTML that did not unfold into readable text. The agent had to use web search as a workaround—relying on search engines to re-index and surface Meilisearch's own content in a more readable form.