Awesome Claude Managed Agents

A fully managed, server-side agent runtime by Anthropic for building and deploying cloud-hosted AI agents at scale. Define an agent, an environment, and start a session — Anthropic handles the agent loop, sandboxed containers, tool execution, and event streaming.

Contents

• Official Resources
• SDKs
• CLI
• Documentation
• Tutorials and Guides
• Community Projects
• Integrations
• Articles and Press
• Videos and Talks
• Related Products

Official Resources

• Product Announcement - Official blog post introducing Claude Managed Agents.
• Engineering Deep-Dive - Architecture post on decoupling the brain from the hands.
• Pricing - Standard token rates plus $0.08 per session-hour metered to the millisecond.
• Research Preview Access Form - Request access to Outcomes, Multi-agent, and Memory features.

SDKs

All SDKs expose Managed Agents under client.beta.agents, client.beta.environments, and client.beta.sessions.

• Python SDK - Install with pip install anthropic, requires version 0.92.0 or later.
• TypeScript SDK - Install with npm install @anthropic-ai/sdk, requires version 0.86.0 or later.
• Go SDK - Install with go get github.com/anthropics/anthropic-sdk-go, requires version 1.33.0 or later.
• Java SDK - Available via Maven or Gradle with com.anthropic:anthropic-java.
• C# SDK - Install with dotnet add package Anthropic.
• Ruby SDK - Install with bundle add anthropic.
• PHP SDK - Install with composer require anthropic-ai/sdk.

CLI

• Anthropic CLI - Go-based ant CLI for managing agents, environments, and sessions from the terminal.

Documentation

• Overview - What Managed Agents are and when to use them.
• Quickstart - Create an agent, environment, and session in minutes.
• Agent Setup - Configure model, system prompt, tools, MCP servers, and skills.
• Sessions - Session lifecycle and status management.
• Environments - Container configuration, packages, and networking modes.
• Tools - Built-in tools, custom tools, and MCP toolsets.
• Events and Streaming - SSE-based event protocol for real-time communication with sessions.
• Cloud Containers - Container specs including Ubuntu 22.04, 8 GB RAM, and 10 GB disk.
• Skills - Reusable filesystem-based expertise packages for domain-specific workflows.
• MCP Connector - Remote MCP server integration with vault-based authentication.
• Permission Policies - Configure always_allow and always_ask tool execution policies.
• Multi-agent - Coordinator agents delegating to callable agents in isolated threads (research preview).
• Memory - Persistent memory stores that survive across sessions (research preview).
• Outcomes - Self-evaluation with rubrics and iterative improvement (research preview).
• API Reference - Full API specification for all endpoints.

Tutorials and Guides

• Verdent Developer Guide - Comprehensive walkthrough for developers getting started.
• Honest Pros and Cons - Practical evaluation with setup instructions.
• Features, Pricing and Early Adopters - Overview of capabilities with early adopter case studies.
• Which Should You Use? - Decision guide for choosing between Managed Agents and Agent SDK.
• Managed Agents vs Agent SDK - Detailed comparison of both approaches.
• Agent Platform Comparison - Comparison of Claude Managed Agents, OpenAI, and self-hosted alternatives.
• Build Your First Claude Agent in 15 Minutes - Step-by-step tutorial with complete Python and TypeScript code walking through agent creation, environment setup, session launch, and event streaming.
• Everything You Need to Know - Comprehensive guide covering architecture, built-in tools, pricing, early adopter case studies, and when to use Managed Agents versus the Messages API or Agent SDK.
• Deploy Your First Production Agent in 10 Minutes - Hands-on walkthrough with practical code and production examples from Sentry, Rakuten, and Asana.
• Claude Managed Agents + Azure: The Multi-Cloud AI Strategy - Analysis of integrating Managed Agents into Azure-first enterprise stacks, covering multi-cloud architectural patterns from a Microsoft Azure engineering perspective.
• Python Tutorial with Code and OpenAI Comparison - Complete Python walkthrough with four working examples covering ClaudeSDKClient, custom tools with the @tool decorator, OAuth credential vault for Slack, and a technical comparison with the OpenAI Responses API.
• Trying Claude Managed Agents for Building Agents at Scale - Developer walkthrough of the Console UI from agent creation to session launch, including a practical GitHub pull request demo using GitHub MCP server authentication via Vault.
• Build a Claude Managed Agent on Vercel - Official Vercel guide for building an internal knowledge agent with per-user credential vaults, durable SSE polling via Vercel Workflows, and a Next.js chat UI connecting GitHub, Notion, and Slack.
• Claude Managed Agents vs Amazon Bedrock AgentCore - Side-by-side comparison of both agent-as-a-service platforms covering architecture philosophy, pricing model differences, and a clear decision framework for choosing between them.
• I Built a Claude Managed Agent in 30 Minutes - Practical build demo connecting a feedback agent to Notion and a live website, with five enterprise case studies covering Notion, Rakuten, Asana, Sentry, and Vibecode.
• Build, Stream, Test: Your First Claude Managed Agent in 30 Minutes - Hands-on Python tutorial building a retail competitive price monitor with real-time SSE streaming and output grading using the Outcomes API.
• Deep Dive: How Claude Managed Agents Solve the AI Scaffolding Nightmare - Technical analysis of the architecture covering credential isolation via dedicated MCP proxies, durable session context management, and the decoupled sandbox model.
• Claude Managed Agents Deep Dive - Comprehensive coverage of architecture, real-world pricing calculations, a Managed Agents vs self-hosted comparison table, and a preview of upcoming features such as private networking and regional deployments.
• Claude Managed Agents Pricing: Costs and Limits (2026) - Detailed pricing breakdown with worked cost calculations, prompt caching analysis, batch API limitations, and guidance on when Managed Agents is cost-effective versus self-hosted alternatives.
• Building a Home Personal Assistant with Claude Managed Agents - Hands-on walkthrough building a Slack-connected household assistant using Managed Agents with Lambda, DynamoDB, and EventBridge Scheduler for async processing and custom memory tools.
• From Building to Deploying: The Complete Guide to Anthropic's Claude Managed Agents - Complete guide covering the conceptual foundation, decoupled sandbox architecture, open-source trade-offs, and hands-on labs building two agents — a Slack support agent and a long-running deep research pipeline — with production hardening patterns.

Community Projects

• CelestoAI/agentor - Open-source alternative inspired by Claude Managed Agents.
• rogeriochaves/open-managed-agents - Self-hostable version with multi-LLM support.
• linear/claude-managed-agents-demo - Integration example connecting Linear with Claude Managed Agents.
• 0xArx/claude-managed-agents-skill - Claude Code skill for building with the Managed Agents API.
• Attilio81/MCP_CMA - MCP server exposing Managed Agents docs to Claude Desktop and Claude Code.
• contro1-hq/centcom-claude-managed-agents - CENTCOM connector for Claude Managed Agents.
• ucsandman/DashClaw - AI agent observability and governance platform with a real-time dashboard, guard policies, audit trail, and an MCP server for integrating with Claude Managed Agents.
• vercel-labs/claude-managed-agents-starter - Official Vercel Labs open-source starter template for an internal knowledge agent connecting GitHub, Notion, and Slack via MCP, built with Next.js 16, Better Auth, Neon Postgres, and Vercel Workflows.

Integrations

• Microsoft Agent Framework - Anthropic provider for the Microsoft agent framework.
• Promptfoo - Evaluation and testing provider for Claude agents.
• Composio - Framework comparison and integration support across agent platforms.

Articles and Press

• SiliconANGLE - Anthropic launches Claude Managed Agents to speed AI agent development.
• The New Stack - Anthropic wants to run your AI agents for you.
• InfoWorld - Anthropic rolls out Claude Managed Agents.
• 9to5Mac - Enterprise features for Claude Cowork and Managed Agents.
• The Register - Anthropic offers to host AI.
• TechRadar - Prototype to launch in days rather than months.
• Help Net Security - Execution and control for AI agent workflows.
• Analytics Insight - Simplifying enterprise AI deployment.
• The Decoder - Managed infrastructure for autonomous AI agents.
• Hacker News Discussion - Community discussion thread from launch day.
• WIRED - In-depth launch coverage including demos from Notion and quotes from Anthropic engineering leadership on the infrastructure challenge Managed Agents solves.
• WinBuzzer - Enterprise-focused analysis covering the OpenClaw ecosystem context, competitive landscape against Amazon Bedrock and Azure AI, and strategic implications of Anthropic's infrastructure push.
• Data Center Knowledge - Industry analysis of how Managed Agents shifts AI workload control to Anthropic's platform, with commentary from Moor Insights and Constellation Research analysts on enterprise lock-in and operational implications.
• Mac Observer - Coverage of Anthropic's simultaneous Managed Agents public beta launch and Claude Cowork general availability, including new enterprise features such as role-based access controls and expanded OpenTelemetry support.

Videos and Talks

• Introducing Claude Managed Agents - Official product overview of the composable APIs including native MCP, tool integrations, memory, and multi-agent coordination features.
• Claude Managed Agents Full Tutorial: How to Setup Your First AI Agent - Hands-on walkthrough building and deploying Claude Managed Agents from scratch, covering sessions, environments, and real-world integrations.
• Claude Managed Agents Full Tutorial & Honest Review - In-depth tutorial and honest review from a developer running 31 scheduled Claude agents, with production insights on costs, patterns, and limitations.
• What is Managed Agents? - Concise explainer of the Managed Agents API covering agents, environments, sessions, and how success criteria drive autonomous execution.
• How Notion Built with Claude Managed Agents - Notion product manager Eric Liu demonstrates delegating complex work to Claude inside a workspace, showing engineers shipping code and teams building websites and presentations using Managed Agents in production.
• Anthropic drops Claude Managed Agents: here's an explanation and demo of what it actually is - Clear explanation and live demo from AI educator Edward Donner clarifying what Managed Agents is, who it is for, and showing a working build with a companion GitHub repository.
• Claude Managed Agents Clearly Explained (and why it matters) - Interview-style breakdown of the four user personas for Managed Agents, real cost data from a live deployment, and a decision framework for when the platform ROI justifies the spend.
• Claude Managed Agents Just Dropped, And It Kills n8n - Practical walkthrough building a transcript-to-ClickUp agent using credential vaults to connect MCP tools without API keys, with a full tour of sessions, the debugging timeline, environments, vaults, analytics, and cost tracking.

Related Products

• Claude Agent SDK - Same engine as Managed Agents, exposed as a library you run on your own infrastructure.
• Claude Code - Agentic coding CLI by Anthropic, built on the same underlying technology.
• Anthropic Skills - Official Claude Code skills repository.

Automated Updates

This list is itself maintained using Claude Managed Agents. A weekly workflow spins up a Managed Agents session with a custom curator skill that searches the web for new resources, scores each candidate against quality criteria, and opens a PR with any additions. See scripts/ for the implementation.

Contributing

Contributions welcome! Read the contribution guidelines first.