A developer wants to build a system where Claude reads and analyzes content from websites. What architecture is required?
Select an answer to reveal the explanation.
Short Explanation and Infographic
Claude can't browse URLs by itself — your code fetches the page, extracts the text, and feeds it to Claude as text in the prompt. Claude analyzes what you give it.
Full explanation below image
Full Explanation
Claude has no built-in web fetching capability in the standard API. For web content analysis, the architecture is: (1) your application fetches the webpage using an HTTP library (requests in Python, fetch in Node.js), (2) extract meaningful text content (using BeautifulSoup, Playwright, Puppeteer, or a content extraction API), (3) include the extracted text in the Claude API prompt, (4) Claude analyzes the text. Some Claude-powered products implement web search as a tool at the product layer, but this is not a base API feature. Option A is wrong — Claude can't fetch URLs. Option C invents a non-existent parameter. Option D is wrong — it's possible, just requires the developer to do the fetching.