Handles user-initiated requests in Perplexity, accessing external content to provide real-time information; not used for automated crawling or AI training.