Anthropic releases Claude Managed Agents for production
The company now offers APIs to build and deploy agents handling infrastructure tasks like session management, state storage, and code isolation. The system splits model logic from execution environment, letting Claude start reasoning before the container spins up.
The Managed Agents architecture eliminates startup delays, cutting median time to first token by 60%. The toolset includes built-in credential stores and the Dreaming system for automated past-session analysis.
Anthropic previously added sandboxing on customer servers and MCP tunneling. Developers can now use pre-built templates in the Claude console or the /claude-api command in Claude Code to set up agents.
Up next
Omegacode: agent-agnostic workflow implementation
Omegacode enables running Claude Code-style workflows with multiple models. You can now combine Claude, Codex, and OpenCode into a single pipeline for adversarial review or deep bug search.
Related stories
- Anthropic releases secure sandboxes for Claude Managed Agents in beta
- Anthropic introduces self-hosted sandboxes and MCP tunnels for agents
- Anthropic hands Agent Skills to the foundation behind MCP
- Anthropic’s daily-brief agent treats a failed read as news
- Claude’s SDKs run the click loop, but the browser is yours
- Opus 5.5 costs less and answers old agent code with 400s
Comments
No comments yet. Be the first.
Join the conversation
Sign in to leave a comment. It appears after moderation.