Cloudflare recently added support for Claude Managed Agents, allowing developers to run and manage Claude agents within Cloudflare. Developers can connect agents to private systems, choose their runtime environment, and monitor agent activity using Cloudflare services.
Cloudflare’s integration with Claude Managed Agents lets developers run agent logic on the Claude platform while using Cloudflare for code execution, secure connectivity, and custom tools. The default deployment includes sandbox controls and logs, support for lightweight or VM-based execution, private service access, browser session monitoring, built-in email capabilities, and customizable tools. Mike Nomitch, director of product at Cloudflare, writes:
Until now, using Claude Managed Agents has meant running the entire stack on Anthropic-provided infrastructure. While this is great for some developers, others may need more control over their infrastructure choice, whether this is for security, compliance, or performance reasons.
Source: Cloudflare blog
As previously reported on InfoQ, Claude Managed Agents let developers run agents on Anthropic's platform with built-in support for browsing, code execution, and tool use. They now support self-hosted sandboxes and MCP tunnels, enabling developers to run agent tools in their own environments while keeping orchestration on Anthropic. Jeremy Daly, founder of Ampt, highlights how Cloudflare is not the only provider taking advantage of the new option:
Cloudflare, Daytona, Modal, and Vercel are all in the launch lineup, with Cloudflare getting its own first-class slot, including Browser Run for automation and quick-start templates.
Anthropic describes this model as "decoupling the brain from the hands." The agent logic and orchestration remain in Anthropic, the "brain," while the execution layer, the "hands," can run on the cloud infrastructure of choice.
The Cloudflare Developer Platform extends agent capabilities with tools such as Git-backed repositories via Artifacts, edge inference with Workers AI, and on-demand application hosting with Dynamic Workers, along with additional infrastructure services. Sebastian Weiss, regional director at Cloudflare, explains why the integration is critical to enterprises and highlights the importance of controlled infrastructure and access:
Picture a bank. The agent’s hands run inside the bank’s own cloud environment. It reaches the core system over a private connection that never touches the public internet. Credentials are injected so the agent itself never holds them. Its actions are logged where the auditors can see them. The brain stays Claude. The control stays with the bank.
Cloudflare provides Mesh and Workers VPC to help connect private services across cloud and on-premises environments without exposing them to the public internet. Claude Managed Agents can use these connections through configurable proxies, allowing teams to restrict agent access, securely inject credentials, and connect only to approved internal services. On Reddit, user boysitisover comments:
Claude and Cloudflare's serverless compute in one product? Sounds like a good way to drain my bank account in about 5 minutes.
Cloudflare provides a customizable control plane as an open-source deployment template, while developers are responsible for managing the costs associated with Claude API usage and Cloudflare resources.