The agent successfully assembled a comprehensive explanation of ADK-TS by combining direct site fetches (homepage, docs structure) with GitHub repository content (README) and the company's main site. The adk.iqai.com website itself is poorly agent-ready—it's a Next.js SPA that renders almost all content via JavaScript, leaving fetches with minimal visible text—so the agent had to rely on GitHub's publicly available README and web search results to answer the core questions. Despite this friction, the agent delivered a complete, accurate answer covering what ADK-TS does, who it's for, pricing (free/open source), and how it differs from alternatives.

• ›Step [13] (GitHub README) was the primary source for technical positioning and core features; it explained ADK-TS as a TypeScript-native AI agent framework with multi-LLM support and type safety. This was the single most information-dense resource.
• ›Steps [10] (GitHub repo page) and [11] (blog post) confirmed the open-source nature and provided context on the project's goals, but the blog post was JavaScript-rendered and yielded minimal text in the fetch.
• ›Step [22] (iqai.com main site) returned Next.js boilerplate with no visible text content, making it unusable for understanding the parent company.
• ›The adk.iqai.com site itself (steps [1], [4], [17]) returned only HTML scaffolding—the agent had to infer documentation structure existed but could not read the actual docs content from the fetches.
• ›Pricing information was not discoverable on adk.iqai.com; the agent inferred it was free/open-source from GitHub metadata and licensing context, not from the official site.
• ›Comparative positioning (vs. LangChain, CrewAI, Google ADK) came entirely from web search results and third-party comparison articles, not from the official adk.iqai.com documentation or marketing.
• ›The site's architecture (Next.js SPA with client-side rendering) is fundamentally hostile to fetched content discovery; all meaningful information required either JavaScript rendering, GitHub fallback, or external search.