LangChain

LangChain is an open-source AI development framework designed for software developers, AI engineers, startups, and enterprise teams building applications around large language models and AI agents. Rather than tying applications to one model provider, LangChain provides standardized interfaces for models, embeddings, tools, retrieval systems, and other components, making it easier to change providers or combine services. Its ecosystem currently includes more than 1,000 integrations covering chat models, embedding models, document loaders, vector stores, tools, and external data sources.

The framework centers on agent development. Developers can use create_agent to build agents that repeatedly call models and tools until a task is completed, while middleware provides customization for guardrails, retries, routing, permissions, context management, and other agent behavior. LangChain agents run on LangGraph, its lower-level orchestration runtime, which provides durable execution, persistence, streaming, human-in-the-loop workflows, and state management. Developers can also use LangGraph directly when they need greater control over complex or deterministic agent workflows.

LangSmith complements the open-source framework with production infrastructure for tracing, debugging, evaluation, monitoring, prompt engineering, sandboxes, and agent deployment. LangSmith can run as a managed cloud service, while Enterprise customers can access BYOC and self-hosted deployment options.

LangChain itself is MIT-licensed and free to use. LangSmith has a Developer plan at $0 per month for one seat with 5,000 base traces monthly. Plus costs $39 per seat per month and includes 10,000 base traces plus one small serverless deployment. Enterprise pricing is custom. Additional traces, compute, storage, deployments, model gateway usage, and other platform services can generate usage-based charges.