Managed Deepagents
A managed AI agent platform introduced by LangChain, enabling agents to autonomously schedule tasks, apply dynamic runtime configurations, and integrate with collaboration tools such as Slack. The platform is designed to allow AI agents to operate as collaborative team members.
Timeline (last 90 days)
The LangChain team released Managed Deepagents version 0.9.
Managed Deepagents v0.9 brings three core updates: autonomous agent schedule creation, per-run dynamic configuration, and Slack message reactions.
The Schedules SDK allows agents to autonomously create reminders, follow-up tasks, and recurring tasks during a conversation.
v0.9 allows defining an agent as a callable function that receives the runtime at the start of each run and returns a define_deep_agent definition, dynamically selecting the model, instructions, skills, MCP servers, and sandbox.
Per-run dynamic configuration determines the configuration before the model runs, so the agent cannot see skills and tools outside the configuration scope, which naturally acts as access control.
The Slack reactions introduced in v0.9 can react with an emoji when the agent takes over a message, with the default being π.
The reactions option allows disabling reactions, replacing the default emoji, or dynamically selecting an emoji for each message.
Managed Deep Agents v0.9 has now entered the Public Beta stage.
The previous version, v0.8, added per-user memory, custom HTTP channels that allow runs to be triggered from any service, and built-in web search capability provided by Parallel.
All Facts (9)
The previous version, v0.8, added per-user memory, custom HTTP channels that allow runs to be triggered from any service, and built-in web search capability provided by Parallel.
50%UnverifiedManaged Deepagents v0.9 brings three core updates: autonomous agent schedule creation, per-run dynamic configuration, and Slack message reactions.
50%UnverifiedThe Schedules SDK allows agents to autonomously create reminders, follow-up tasks, and recurring tasks during a conversation.
50%Unverifiedv0.9 allows defining an agent as a callable function that receives the runtime at the start of each run and returns a define_deep_agent definition, dynamically selecting the model, instructions, skills, MCP servers, and sandbox.
50%UnverifiedPer-run dynamic configuration determines the configuration before the model runs, so the agent cannot see skills and tools outside the configuration scope, which naturally acts as access control.
50%UnverifiedThe Slack reactions introduced in v0.9 can react with an emoji when the agent takes over a message, with the default being π.
50%UnverifiedThe reactions option allows disabling reactions, replacing the default emoji, or dynamically selecting an emoji for each message.
50%UnverifiedManaged Deep Agents v0.9 has now entered the Public Beta stage.
50%UnverifiedThe LangChain team released Managed Deepagents version 0.9.
50%