Build a scraper to understand the structure of websites and as a reference I used my favorite coffeeshop where I get to work on this. For this I had a couple setbacks on the firecrawl mcp part so I would strongly recommend using claude for the trouble shooting while loading screenshots along the way