Wire LangChain retrievers — vector stores, document transformers, embedding models, metadata filtering, hybrid search, reranking, and query rewriting. Use when Codex adjusts retrieval flow inside LangChain applications.
