Anthropic has officially released a public beta version of its Claude Managed Agents. This new infrastructure allows users to configure and deploy complex multiagent workflows, marking a transition from providing access to standalone models toward offering managed agent systems.
The update introduces several technical features designed to support production-level workloads. These include server-evaluated permission policies and the ability to establish live session connections. Anthropic states that this release does not include a new AI model or updated benchmark results.
