What are Retrievers in LangChain and how do they work in RAG applications?

I'm learning LangChain and Retrieval-Augmented Generation (RAG), and I often come across the term Retriever. I understand that retrievers are used to fetch relevant information from a knowledge base, but I'm not sure how they work or how they differ from vector databases and embeddings.

Can someone explain what retrievers are, how they retrieve relevant documents using embeddings and vector stores, and the different types of retrievers available in LangChain? A simple example showing how a retriever works with a vector database like FAISS, ChromaDB, or Pinecone in a RAG application would be very helpful.