LangChain — LLM Application Framework

LangChain is the most popular framework for building LLM applications. It provides: • Abstractions for LLMs, chat models, and embeddings • Prompt templates and output parsers • Chains — sequences of LLM calls and tools • Agents — autonomous decision-making with tools • Memory — short and long-term conversation memory • Document loaders, text splitters, vector stores for RAG

LLM/ChatModel — Unified interface for any LLM provider (OpenAI, Anthropic, Cohere, HuggingFace). PromptTemplate — Parameterized prompt strings. OutputParser — Parse LLM output into structured formats. Chain — Sequence of steps: prompt → LLM → parser → next step. LCEL (LangChain Expression Language) — Pipe syntax: chain = prompt | llm | parser