10h ago

Claude announces the public beta of self-hosted sandboxes and research preview of MCP tunnels within Claude Managed Agents, enabling runs inside user infrastructure under existing security controls

3887.1K5712.9K1.2M

——0——

Features were presented at Code with Claude London with Cloudflare and Modal support.

Original post

#322@BCHERNYOP

Claude@CLAUDEAI

Live from Code with Claude London: we're launching self-hosted sandboxes (public beta) and MCP tunnels (research preview) in Claude Managed Agents. Run agents inside your own perimeter, with your security controls applied by default.

12:57 AM · May 19, 2026

Reposted by

#1511@MATTTURCK

#863@CHARLES_IRL

QUOTE POST

#682Guillermo Rauch@RAUCHG

Claude Managed Agents 🤝 Vercel Sandbox https://vercel.com/kb/guide/run-claude-managed-agent-tools-with-vercel-sandbox

Claude@claudeai

7:57 AM · May 19, 2026 · 943.4K Views

1:57 PM · May 19, 2026 · 16.1K Views

Sentiment

Pos91.1%

91.1%

50%

Neg8.9%

Positive users praise Claude's self-hosted sandboxes, MCP tunnels, and Cloudflare/Vercel integrations for enabling private, production-ready agent deployments, while negative users worry about security risks and unexpected costs.

119 comments with sentiment.

DIGG DEPTH

Kevin Rose@KEVINROSE

How will this work with Vercel?

Vercel Sandbox runs Claude Managed Agents' tool calls (run_shell, read_file, etc.) inside fresh per-session microVMs.[1] Anthropic handles the core model, loop, and state; Vercel supplies the webhook control plane, credential brokering, and deny-by-default network policies.

Developers scaffold a Next.js app with @vercel/sandbox@beta, build a snapshot, and register the webhook for self-hosted execution under their own infrastructure.

Cluster engagement

35 snapshots