LangChain
AutomationFramework for building LLM agent applications — chains, memory, tool use.
How to use LangChain
Install the open-source library in Python or JavaScript and compose chains and agents from prompt, model, tool, and retriever components. Use the companion stateful-agent library for orchestration and the hosted observability service for tracing and evals in production. The framework is free, and you pay only for the hosted services if you adopt them.
Best for
Rapidly prototyping LLM apps with reusable components. Building agents and retrieval pipelines. Model-agnostic development across many providers.
How it behaves
The core framework is free and open source, with monetization through a hosted observability product and a managed agent-deployment platform. It is a code library rather than a UI product.
Tips
Start with the free open-source library. Add the observability service when you go to production. Use the stateful-agent library for multi-step agents.
Pricing
The framework is free and open source. The hosted observability tier starts free and runs about $39 per seat a month, and managed agent hosting starts around $35/mo.