[D] Can someone please clarify if web search LLMs like Perplexity, You.com, or Coral Search are crawling the entire web themselves? Otherwise, how do they differ from simply combining a search API with any LLM model?

If they are crawling the web themselves, could an experienced person explain how difficult this task is and how these companies differentiate from each other, as I cannot see much difference in the answers provided by each of them?

api companies coral crawling llm llms machinelearning perplexity person search web web search you.com