The agent successfully assembled a comprehensive overview of Journalit by combining information from the homepage, comparison pages, enterprise page, and GitHub README—but had to work around a fragmented site structure and heavy client-side rendering that obscured pricing and key details. The site does not maintain a dedicated, fetchable pricing page (/pricing returned 404), forcing the agent to reconstruct pricing from meta tags, search results, and scattered pages. Core identity information (what it does, who it's for, how it differs) was ultimately discoverable but required cross-referencing multiple sources rather than being centralized.

• ›The homepage [1] was rendered client-side JavaScript with truncated HTML, so pricing and feature details were not directly extractable from the raw fetch; the agent could not read the pricing information from the initial fetch and had to resort to web search and alternative pages.
• ›Pricing information was found embedded in meta descriptions and scattered across pages like /enterprise [10] and /compare [9], but no dedicated /pricing page exists (returned 404 in step [2]). The agent had to infer pricing from multiple fragments—$18/month and $180/year emerged from search snippets and the homepage meta tag, not from a primary source.
• ›The GitHub repository [13] and README [14] provided authoritative product description and confirmed the local-first, Obsidian-plugin nature of the product, but this external source was necessary because the site's own documentation was fragmented.
• ›Comparison pages [9, 11] were the most navigable—they directly addressed 'how it's different from alternatives' and were explicitly findable via web search. The site's competitive positioning is strongest on these dedicated comparison pages rather than on a central features or about page.
• ›The agent correctly identified that core journaling is free but struggled to clearly determine what features require authentication or payment within the free tier, indicating the site does not clearly separate capability boundaries in its public documentation.
• ›The site is agent-unfriendly for structured discovery: no /pricing, /features, or /about pages; no sitemap or robots.txt hints; heavy JavaScript rendering; information scattered across marketing pages, comparison pages, blog posts, and external GitHub. The agent succeeded by combining prior knowledge, web search routing, and manual page-hopping rather than following a designed information architecture.