langchain-ai/langchain

LangChain is a framework that helps developers build applications powered by large language models, also called LLMs (the AI systems behind chatbots like ChatGPT). The core idea is that real-world AI applications usually need to chain several pieces together, like a model, a data source, a tool, and some logic, and LangChain gives you a standard way to wire those pieces up so you can swap any of them later. In practice, you install the LangChain package and write a few lines of code that pick a chat model, send it a prompt, and get a response back. From there, you can add components like retrievers that pull in your own data, tools the model can call, and integrations with outside services. Because everything follows a common interface, you can change the underlying model, vector store, or provider without rewriting your whole application. The project also points users toward sister tools in the same ecosystem: Deep Agents for higher-level agent patterns like planning and subagents, LangGraph for more controlled agent workflows, and LangSmith for evaluating, monitoring, and debugging agents in production. You would reach for LangChain when you are building something more involved than a single chat call, for example an assistant that searches your documents, an agent that uses tools, or any app where you want the option to switch models and providers later. The framework is written in Python and there is a separate JavaScript and TypeScript version called LangChain.js for web and Node projects.