LangChain
Open-source framework for building applications with language models; simplifies chaining prompts, tools, and memory.
Think of it like
The glue that makes LLM apps easy; handles the plumbing so you focus on logic.
Example
Load a vector store, prompt templates, tool definitions, then chain them: "retrieve docs → summarize → generate answer."
How it actually works
LangChain provides abstractions for common patterns: retrievers, memory, agents, output parsers. Works with any LLM via a standard interface. Hugely popular; has spawned competitors (DSPy, Langraph).
For product teams
Market leader in LLM orchestration; building payment/hosting layer (LangSmith).
For engineers
Simplifies building RAG, agentic systems, and complex chains. Abstraction over APIs means switching models is easy.
Read anything AI without the jargon
Look up any term in plain English, or save terms as you read with the free Chrome extension.
Open DecoderAdd to Chrome