Docker has announced Docker Cloud Sandboxes, a new solution for secure, isolated AI agent execution that enables complex agentic workflows to continue running in the cloud long after a developer’s laptop shuts down. Launched at WeAreDevelopers North America, Docker Cloud Sandboxes let organizations run agentic workloads at scale without tying up developers’ hardware, provisioning their own infrastructure, or paying for unused capacity.
Docker Cloud Sandboxes give developers a straightforward path to move their agentic workflows to the cloud, whether extending an existing local sandbox setup or starting a new build.
As agentic workflows mature, AI agents are running longer, working unattended, and taking on more complex, higher-stakes work that increasingly needs to continue whether or not a developer is at their machine. Developers need the freedom to run any agent on any model and the ability to move execution off individual laptops without giving up the isolation and control that make agents safe to run.
Docker Cloud Sandboxes extend Docker’s secure local sandbox isolation into the cloud, giving developers the flexibility to run AI agents wherever it makes sense for the workflow, without having to make a binary choice between running locally or in the cloud.
“Modern agentic workflows necessitate a fundamentally different approach to isolation,” said Mark Cavage, President at Docker.
“While the containers Docker is known for still have a critical role to play, they weren’t designed for the level of isolation AI agents demand. That’s why we introduced local sandboxes as a new, separate solution earlier this year, and why we’re now extending the same level of isolation and control to the cloud. With Docker, developers get the same sandbox with the same policies whether they run their workflows locally or in Docker’s cloud, enabling them to unleash agents on any task, no matter its scope or stakes.”
Extending isolation from laptop to cloud
Docker Cloud Sandboxes build on the foundation of the company’s local sandboxes, delivering the same CLI, trust model, and Kits, enhanced by elastic cloud capacity. With Docker Cloud Sandboxes, developers can:
- Turn off the laptop and keep the work going. Agents run in Docker-managed cloud infrastructure rather than on a developer’s machine, with the same microVM isolation and policies as local sandboxes.
- Start locally, then shift the work to the cloud. Workflows scale out from a laptop or the web, without ever having to switch environments.
- Execute long jobs overnight. Agents can work unattended for hours without tying up a developer’s machine.
Boot up in low hundreds of milliseconds. Docker Cloud Sandboxes are ready to go instantly, with secrets, policy, MCP gateways, and agent configuration already built in. - Scale up parallel tasks, with no infrastructure to provision. Compute scales from 1 to 16 vCPUs and is fully managed by Docker; developers just point their agents at it.
- Build agent environments once, run them anywhere. Developers write their commands once and run them either locally or in the cloud, under the same controls, with no changes required.
Together with its AI Governance solution, Docker’s local and cloud sandboxes give organizations the baseline they need to bring greater control, security, and reliability to the environments where agents execute.
“Nobody needs convincing anymore that AI can drive real productivity gains,” said Mat Velloso, Chief Product Officer at Docker.
“The problem that’s still largely unsolved is how to achieve those gains safely and with the right guardrails. Every organization that can’t unleash agents at scale without something breaking is leaving real value on the table. That’s exactly what we’re here to solve, and it all starts with isolation and deterministic policies. Docker’s sandboxes deliver a secure, model- and harness-neutral environment for running any agentic workload, locally or in the cloud, with policies that the agent can’t ignore or change.”
Introducing next-generation Kits: An open standard for packaging agentic sandboxes
Docker also published the next generation of Kits, the company’s open specification for packaging an agentic sandbox, including the agent, its tools, and the rules for what it can touch, as a single, shareable artifact.
Kits are now built as standard OCI images, the same open standard behind every container, so there’s no proprietary format and no vendor lock-in. Developers can build and ship Kits with tools they already know, quickly and at low cost, using the same process they already use to publish container images on Docker Hub.
Access rules travel inside the Kit itself, so enterprises can define what an agent can touch once and have it enforced everywhere the Kit runs, solving the loss of control that has historically held them back from deploying agents. In addition, because Kits snap together from mixins, enterprises can standardize proven parts while letting teams fit the rest to their own workflows.
Docker has committed to submitting its Kits specification to the Cloud Native Computing Foundation (CNCF), the open source, vendor-neutral hub of cloud-native computing.
“Standards are what let an ecosystem move fast without fragmenting, and few companies understand that better than Docker through their involvement in efforts like the OCI and CNCF,” said Chris Aniszczyk, CTO at CNCF.
“By delivering Sandbox Kits as standard OCI images, Docker is giving the industry an open, repeatable way to package an AI agent, its tools, and its guardrails as one artifact. OCI is the foundation the cloud native ecosystem is built on, so a standard for agents that builds on OCI reaches the whole ecosystem at once. The cloud native community looks forward to working with Docker to bring this work under neutral governance,” Aniszczyk concluded.