JAKARTA - Cloudflare, Inc. is expanding its Agent Cloud with a variety of new features to help developers build, deploy, and scale agents.
This suite of infrastructure, security, and developer tools is designed to move AI agents from experimental demos on local laptops to robust production-level workloads running across Cloudflare's global network.
Through these new features, Cloudflare aims to handle infrastructure, computing, deployment, and security so that developers can focus on building next-generation applications.
"We're entering a world where agents become writers and run code. However, agents require a default secure environment, can be scaled to millions instantly, and still run even if the task lasts a long time," said Matthew Prince, co-founder and CEO of Cloudflare.
One of the new features introduced is Dynamic Workers, an isolation-based runtime designed to execute AI-generated code quickly and efficiently.
This technology allows agents to perform tasks in milliseconds without warm-up time, at a lower cost than traditional containers.
Cloudflare also introduced Artifacts, a Git-compatible storage system designed for the era of AI agents. This feature allows the creation of millions of repositories, remote source forks, as well as providing permanent space for agents to store code and data.
In addition, the Linux-based Sandbox is now generally available to run tasks that require a full operating system, where agents can clone repositories, install Python packages, run builds, and iterate with the same feedback cycle as human developers.
To support long-term tasks, Cloudflare introduces Think in the Agents SDK, which allows agents to run multi-step and persistent jobs.
Powered by GPT-5.4 and Codex to run real enterprise workloads at scale, Cloudflare's Agent Cloud is expected to make it easier for developers to deploy production-ready agents.
The English, Chinese, Japanese, Arabic, and French versions are automatically generated by the AI. So there may still be inaccuracies in translating, please always see Indonesian as our main language. (system supported by DigitalSiber.id)
