The open source stack for building agents

Deep Agents gives you a production-ready harness for building LLM-powered agents and applications. Deep Agents ships with built-in context management and abstractions for subagents, long term memory, and skills. Deep Agents are built to run reliably for long-running, large context tasks.

^{Use Deep Agents when you need to:}‍

• Give agents memory, tools, skills, and filesystem access
• Delegate work to subagents with isolated context
• Manage long context without building summarization and offloading yourself
• Customize a powerful base harness without starting from scratch

LangChain is the framework for building agents. It provides create_agent, a core agent loop built on LangGraph, plus the building blocks for models, inference providers, tools, messages, MCP, and middleware. Use LangChain when you want to assemble your own agent harness.

^{Use LangChain when you need to:}

• Build around the standard agent loop with a model and tools
• Add middleware around the agent loop
• Inject guardrails, dynamic context, human review, or business logic
• Build a bespoke harness with your own defaults and controls

LangGraph is the runtime for custom agent workflows. It uses a graph-based model backed by a durable engine, with human-in-the-loop control, fault tolerance, streaming, persistence, and observability at every step. Use LangGraph when your agent doesn’t fit a standard loop, or when you need to combine deterministic and agentic steps in the same workflow.

^{Use LangGraph when you need to:}‍

• Mix deterministic steps with agentic behavior
• Add custom approvals, interrupts, retries, and fault tolerance
• Model complex state transitions directly
• Build multi-agent systems with explicit coordination
• Control how every step of the agent runs