The community members are discussing how to build a search engine like Perplexity. One community member suggests that Perplexity likely uses Bing Search under the hood, and that it could be built using a tool like LlamaHub. Another community member wonders how Perplexity synthesizes answers so well, and mentions that it may involve more than just Retrieval Augmented Generation (RAG). A third community member references a video that discusses Perplexity's use of multiple LLM calls to provide a rich response. However, there is no explicitly marked answer in the comments.
I'm pretty sure Perplexity just uses Bing Search under the hood. They probably do some other processing but it's mostly just a search API connected to a LLM.
