[How to: return sources](https://python.langchain.com/docs/how_to/qa_sources/): LLM should read this page when: 1) Building a question-answering (QA) application that needs to return the sources used to generate the answer. 2) Implementing a conversational QA system with retrieval-augmented generation (RAG). 3) Structuring model outputs to include sources or citations. 'This guide explains how to configure LangChain's QA and RAG workflows to retrieve and return the source documents or citations used to generate the final answer. It covers both basic RAG and conversational RAG architectures, and demonstrates techniques for structuring the model output to include source information.'

