The agent successfully assembled a comprehensive explanation of assistant-ui by combining direct site fetches (which were heavily JavaScript-rendered and unreadable), GitHub exploration, and third-party review aggregators. The official website proved nearly opaque to text extraction—pricing and docs pages returned only boilerplate HTML—forcing the agent to rely heavily on search-indexed third-party sources (DesignRevision, AIAgentsDirectory) and GitHub to answer the core questions about what the product does, who it's for, and how it's priced. The final response is accurate and complete, but highlights a significant agent-readiness gap: the site does not publish its key positioning, feature list, or pricing in a machine-readable or fetchable form.

• ›The assistant-ui.com homepage, pricing page, and docs pages all returned only HTML boilerplate with Next.js preload directives—no visible content. The site is client-side rendered with no server-side content, making it completely inaccessible to text-based agents.
• ›The agent found pricing details (Free tier: 200 MAU, Growth: $50/month, Enterprise: custom) only through third-party review sites (AIAgentsDirectory, designrevision.com) and search results, not from the official pricing page itself.
• ›GitHub (step [9]) and the DesignRevision alternatives comparison (steps [8], [19]) were the only content-bearing external sources. These provided feature descriptions, positioning relative to competitors (CopilotKit, Vercel AI SDK, NLUX), and the architecture insight about 'headless components' and 'composable UI'.
• ›The agent explicitly called out what it could not access: 'The pricing page doesn't display in my text scraper—It's heavily client-side rendered' and 'Unclear exact scope of cloud features.' These gaps required the agent to note ambiguities rather than deliver complete coverage.
• ›Search results (steps [6], [7], [11], [14], [15], [18], [21]) were essential routing—they pointed to external review sites and GitHub—but the search snippets themselves did not contain full answers. The agent had to fetch the downstream pages.
• ›The site's own documentation link to 'Headless Chat Primitives' (found in search results at step [18]) was cited in sources but the agent could not fetch it directly due to client-side rendering, so the insight about composable architecture likely came from prior knowledge or DesignRevision's comparison article.