Skip to main content
Deep Agents is LangChain’s agent harness for long-running agents with planning, filesystem access, subagents, memory, and tool use. Braintrust traces Deep Agents runs so you can inspect the agent workflow, tool calls, and model calls in one trace.