Anthropic has launched Claude Managed Agents, a cloud-hosted platform that enables development teams to deploy AI agents at production scale without building or managing infrastructure.

The system provides managed execution for sandboxing, state management, authentication, and error recovery, moving the operational burden of agent deployment from internal engineering teams to Anthropic's infrastructure.

This matters because AI teams are increasingly focused on moving agents from prototype to production — and infrastructure complexity has been the primary blocker to that transition.

According to Anthropic, Claude Managed Agents is designed to reduce the time required to move from prototype to production — from months to days — by combining a built-in orchestration system with managed infrastructure optimized for Claude models.

The launch directly impacts enterprise developers, AI engineering teams, and product leaders building agent-based workflows, particularly those constrained by infrastructure cost, reliability, or governance requirements.

In short, Claude Managed Agents is a fully managed agent runtime that allows teams to define agent behavior while Anthropic handles execution, scaling, and operational complexity.

Claude Managed Agents is a cloud-hosted AI agent platform that abstracts infrastructure, orchestration, and execution so developers can focus on building agent behavior instead of managing systems.

Claude Managed Agents is Anthropic's managed AI agent platform designed to remove infrastructure barriers and accelerate enterprise deployment of production-ready agents.

Anthropic launched Claude Managed Agents as a public beta product available through the Claude Platform and Claude Console. The offering is a suite of composable APIs — meaning individual components can be selected, combined, and configured to match specific agent workflows, rather than requiring teams to adopt a single fixed architecture.

The launch addresses one of the most persistent problems in enterprise AI development. Getting a production agent off the ground has historically meant building out secure execution environments, managing state and checkpointing, handling credentials and permissions, and wiring up end-to-end visibility — all before a single user interaction. That engineering burden alone has kept production agents out of reach for many teams. Claude Managed Agents consolidates all of it into a performance-tuned orchestration layer running on Anthropic-operated infrastructure, taking teams from prototype to launch in days rather than months. Development teams define the agent's tasks, tools, and guardrails — the orchestration layer handles when to call tools, how to manage context, and how to recover from errors.

Claude Managed Agents is available now on the Claude Platform, accessible through the Claude Console or the new CLI for terminal-based deployment. Developers using the latest version of Claude Code can also build with Managed Agents directly — just prompt "start onboarding for managed agents in Claude API" within the built-in claude-api Skill to get started. Teams access the platform on a consumption model, paying standard Claude Platform token rates plus $0.08 per session-hour for active agent runtime

The core of Claude Managed Agents is a built-in orchestration layer that determines when to invoke tools, how to manage context across a session, and how to recover from errors mid-task. Rather than requiring developers to write and maintain this logic themselves, Managed Agents runs it on Anthropic's infrastructure, already tuned for Claude model performance.

The platform includes 4 primary capability areas:

With Managed Agents, development teams set the desired outcomes and define success criteria upfront — Claude then runs its own evaluation process and continues refining until those targets are met. This capability is currently in research preview and requires an access request through Anthropic. For teams that prefer more direct control, the platform also supports standard prompt-and-response workflows. Anthropic's internal testing on structured file generation found that Managed Agents outperformed standard prompting loops by up to 10 points on task success, with the biggest improvements showing up on the most complex tasks.

All of that activity is visible inside the Claude Console where developers can review every tool call, trace decisions, and diagnose failure points without standing up a separate monitoring system.

Several enterprise teams were already building and shipping with Claude Managed Agents at launch, with use cases spanning productivity, engineering, finance, and sales.

Notion integrated Managed Agents into its workspace to support Notion Custom Agents, currently in private alpha. The integration enables users to delegate complex, open-ended tasks — coding, generating presentations, building spreadsheets, and producing websites — without leaving the Notion environment.

Eric Liu, Product Manager at Notion, described the integration this way: "We want Notion to be the best place for teams to work with agents and get things done. We integrated Claude Managed Agents, which can handle long-running sessions, manage memory, and deliver high-quality outputs over time, to make that possible. Our users can now delegate open-ended, complex tasks, everything from coding to generating slides and spreadsheets, without ever leaving Notion."

Rakuten deployed specialist agents across product, sales, marketing, and finance — integrated directly into Slack and Microsoft Teams so employees can assign tasks and receive deliverables including spreadsheets, slides, and apps. Each specialist agent was deployed within a single week.

Yusuke Kaji, General Manager of AI for Business at Rakuten, said: "With Claude Managed Agents, our power users become like Galileo, contributing across domains far beyond a single specialty or discipline. We deploy each specialist agent within a week, managing long-running tasks across engineering, product, sales, marketing, and finance, generating apps, proposal decks, and spreadsheets in sandboxed environments. As agents become more capable, Managed Agents lets us scale safely without building agentic infrastructure ourselves, so we can focus entirely on democratizing innovation across the company."

Asana used Managed Agents to build AI Teammates — collaborative AI agents that work alongside human team members inside Asana projects, picking up tasks and drafting deliverables. According to Anthropic, the Asana team used Managed Agents to add advanced features significantly faster than would have otherwise been possible.

Sentry paired its existing Seer debugging agent with a Claude-powered agent that writes the code fix and opens a pull request (PR), creating an end-to-end flow from flagged bug to reviewable fix.

Indragie Karunaratne, Senior Director of Engineering, AI/ML at Sentry, described the decision: "Turns out telling developers what's wrong with their code isn't enough: they want you to fix it too. Customers can now go from Seer's root cause analysis straight to a Claude-powered agent that writes the fix and opens a PR. We chose Claude Managed Agents because it gives us a secure, fully managed agent runtime, allowing us to focus our efforts on building a seamless developer experience around the handoff. Managed Agents not only allowed us to build the initial integration in weeks instead of months, but has also eliminated the ongoing operational overhead of maintaining bespoke agent infrastructure."

Vibecode adopted Managed Agents as its default integration for going from prompt to deployed application, enabling a new generation of AI-native app development. According to Anthropic, Vibecode users can now spin up that infrastructure at least 10 times faster than before.

The range of industries and use cases at launch suggests Claude Managed Agents is not solving a niche developer problem. It's addressing a deployment bottleneck that cuts across the entire enterprise AI landscape.

The speed advantage matters beyond developer convenience. Every week of infrastructure work delayed before an agent reaches users is a week of value not delivered. By moving infrastructure responsibility from the development team to Anthropic's platform, teams redirect engineering cycles toward product decisions: what the agent should do, how it should handle edge cases, and what guardrails it operates under, rather than toward the machinery required to run it.

For enterprise teams operating in environments that require auditability and access control, the trusted governance layer is a meaningful enabler. Scoped permissions limit what systems an agent can touch. Identity management ensures agents act under defined credentials. Execution tracing produces a full record of every decision an agent made, a requirement in regulated industries and a strong signal for AI governance teams evaluating deployment readiness.

For teams without dedicated infrastructure engineers or the budget to build agentic systems from scratch, Claude Managed Agents effectively removes the technical barrier that has kept production AI agents out of reach.

Q: What is Claude Managed Agents and what did Anthropic launch?
A: Claude Managed Agents is a cloud-hosted AI agent platform launched by Anthropic in public beta. It provides development teams with fully managed infrastructure for building and deploying AI agents, including sandboxing, orchestration, long-running sessions, and observability.

Q: How does Claude Managed Agents work?
A: The Key Point: Claude Managed Agents removes the need for teams to build infrastructure by running agent orchestration, execution, and recovery directly on Anthropic's platform. Developers define tasks and tools, while the system handles execution, state management, and scaling.

Q: Why does Claude Managed Agents matter now?
A: AI teams have struggled to move agents from prototype to production due to infrastructure complexity. Claude Managed Agents reduces that barrier, enabling teams to deploy production agents in days instead of months.

Q: What are the current limitations?
A: Multi-agent coordination and self-evaluation features are currently in research preview and require access requests. The platform is otherwise available in public beta.

Q: How is Claude Managed Agents priced?
A: Pricing is usage-based, including standard Claude token costs plus $0.08 per session-hour for active runtime.

Q: Which companies are using it?
A: Notion, Rakuten, Asana, Sentry, Vibecode, and others are already deploying agents in production using the platform.

The launch of Claude Managed Agents shows that enterprise AI competition is expanding beyond model performance and into the infrastructure required to deploy agents reliably in production.

Key point: Anthropic is expanding from AI models into infrastructure, providing the execution layer required to run agents reliably in production environments.

Who should care: Enterprise developers, AI engineering teams, and product leaders building agent-based systems should evaluate this launch closely, particularly those limited by infrastructure complexity or deployment timelines.

Why this matters now: The gap between AI prototypes and production systems has consistently slowed enterprise adoption. By moving infrastructure responsibility to the platform layer, Claude Managed Agents directly addresses one of the most persistent barriers to deployment.

What decision this affects: Teams deciding how to deploy AI agents now face a clear choice: build and maintain custom infrastructure or adopt a managed platform with faster deployment timelines and reduced operational overhead.

In short, Claude Managed Agents removes the infrastructure barrier that has historically slowed AI agent adoption, allowing teams to focus on product design, workflows, and real-world use cases instead of system maintenance.

Anthropic is not just making agents easier to build — it is making them far harder for enterprise teams to ignore.

Editor’s Note: This article was created by Alicia Shapiro, CMO of AiNews.com, with writing support from Claude, and AEO/GEO/SEO optimization, image concept development, and editorial structuring support from ChatGPT, AI assistants. All final editorial decisions, perspectives, and publishing choices were made by Alicia Shapiro.