Cloudflare partners with Anthropic to launch secure sandboxes for AI agents

Cloudflare Inc. (NYSE: NET) announced a collaboration with Anthropic to launch Cloudflare Environments for Claude Managed Agents, according to a company press release. The integration allows organizations to run agent operations on Claude's platform while using Cloudflare's network and Workers platform for code execution and security.

The service provides secure sandboxes for AI agent sessions through a Workers-based control system. Organizations can scale from single prototypes to millions of users while maintaining security protocols, the company stated.

"The true potential of AI agents will only be realized when they can safely interact with the real world at massive scale," said Matthew Prince, co-founder and CEO of Cloudflare.

The platform offers multiple runtime options including Linux-based microVMs for complex tasks and V8 Isolate-based sandboxes with millisecond boot times. The system includes Zero-Trust connectivity features to protect sensitive data through encrypted connections to internal services.

Additional features include built-in audit trails and session recordings for compliance purposes, along with the ability for developers to write custom functions directly on Cloudflare's platform to extend agent capabilities.

The launch represents the latest development in the ongoing partnership between Cloudflare and Anthropic focused on AI connectivity infrastructure. Cloudflare operates one of the world's largest interconnected networks and serves millions of organizations globally.

You May Also Be Interested In

• DOJ Unveils $6.5B Healthcare Fraud Crackdown - WSJ
• Cloudflare and beehiiv partner on AI crawler controls for publishers
• Zeta Surgical begins TMS navigation treatments at Hope Therapeutics clinics