Managed Deep Agents

If you're building agents with LangSmith and want to skip the server setup, this is the deployment path. You define your agent in a local directory (Python or TypeScript), write tools and instructions as normal files, then run `mda deploy` to push it to managed infrastructure. The runtime handles durable execution, sandboxes, memory, and traces. You get Context Hub for instructions, MCP connectors for external tools, and cron schedules if you need background work. During private beta it's CLI-only and US Cloud only. Use standard LangGraph deploys if you need custom routes, advanced auth, or more control over the stack.

skills.sh installs: 10K
GitHub stars: 1.3K
last updated: Aug 21, 2026

View on GitHub

Install

$ npx -y skills add langchain-ai/langchain-skills --skill managed-deep-agents --agent claude-code

Installs into .claude/skills of the current project.

Files

Managed Deep Agents

Overview

Managed Deep Agents (MDA) is a hosted runtime for code-first Deep Agents in LangSmith. You author an agent in Python or TypeScript, test it locally with mda dev, and deploy it with mda deploy. It runs the open-source Deep Agents harness (see [[deep-agents-core]]) on LangSmith Agent Server, and adds managed sandboxes, browsers, memory, connections, schedules, and channels as project-file declarations.

The core idea is that an agent is a directory. A file's location determines its role, and the CLI compiles that directory into a managed LangGraph app.

MDA is in public beta on LangSmith Cloud and BYOC data planes.

The docs are the source of truth

This skill stays thin on purpose. It tells you what to build and where it goes; the docs tell you exactly how. Before writing a capability's file, fetch its docs page from the feature map and follow it.

SKILL.md on GitHub