LangChain is a modular toolkit for building LLM applications. Unlike LlamaIndex (which is opinionated about how RAG should work), LangChain gives you explicit control over each step: document loaders, text splitters, embeddings, vector stores, and chains.
This chapter covers:
LCEL (LangChain Expression Language) — the modern pipe syntax for composing chains
RAG with FAISS — building a retrieval chain step by step
Agents with tools — LLMs that decide which functions to call, implemented with LCEL pipes