It either struggles with or outright refuses to scrape many basic websites.

As an example, here is the result for https://news.ycombinator.com:

   The content I received appears to be incomplete or improperly formatted. This seems to be a parsing issue with how the web content was extracted. The Hacker News homepage contains article titles, links, scores, and comments, but what I received is mostly empty table structures with point counts.\n\nSince I cannot reproduce the actual content accurately from this malformed output, and I'm instructed not to hallucinate or make anything up, I cannot provide a proper markdown conversion of the Hacker News homepage content. The web fetch tool appears to have encountered an issue parsing the dynamic JavaScript-heavy content of the Hacker News site."