Anthropic releases secure sandboxes for Claude Managed Agents in beta
Anthropic has launched public beta testing of self-hosted sandboxes and research previews of MCP tunnels for Claude Managed Agents. These tools enable large businesses to run agents within their own secure perimeter.
Companies can now deploy isolated runtime environments on their own servers or through providers like Cloudflare, Daytona, Modal, and Vercel. This addresses security concerns when working with sensitive data and heavy computations.
Key technical updates include support for long-lived sessions, Zero Trust secret management, and secure outbound network connections for MCP tunnels. The latter allows agents to access internal services without publishing endpoints to the internet.
Up next
AgentPTY: bypassing Claude Code subscription limits
Developers found a way to automate Claude Code for free, bypassing Anthropic’s upcoming paid limits. The new AgentPTY tool launches the console interface inside a virtual terminal and controls it through a local API. To Anthropic’s servers, this looks like regular human activity, saving on developer subscription costs.
Related stories
- Anthropic releases Claude Managed Agents for production
- Anthropic introduces self-hosted sandboxes and MCP tunnels for agents
- Claude Duet won’t share your card number even if you ask
- Claude’s auto mode skips connector “Always ask” settings
- Anthropic’s test agents filled out US visa applications
- Claude desktop teardown: MCP servers run outside the VM
Comments
No comments yet. Be the first.
Join the conversation
Sign in to leave a comment. It appears after moderation.