LangChain adds and revamps async subagents in Deep Agents

LangChain adds async subagents to Deep Agents, so a background worker can run a long task while the orchestrator stays available to the user.

Making a subagent async takes one field, "async": True, in its definition. Async subagents are available in Deep Agents for Python and JavaScript, and in Managed Deep Agents, where every served agent also gets an async copy of the general-purpose subagent that can run delegated work in either the foreground or the background. LangChain says task notifications are in beta and the API may still change.

On November 5, LangChain posted on X that it revamped support for async subagents in Managed Deep Agents. The post gives no details of the changes.