The Pulse
OpenClaw Announces Enterprise Control Plane for Persistent Agents
OpenClaw announced an open-source control plane for managing persistent agents in sensitive environments, with support for multiple tenants, security boundaries and lifecycle governance. The project is available for internal pilot workloads
AI.info Team · Written with AI ·
OpenClaw announced OpenClaw Enterprise (OCE) on September 29, describing it as an open-source, vendor-neutral platform for managing persistent agents in sensitive environments. The project adds multi-tenancy, hard security boundaries and standardized agent primitives, with governance and auditability across the agent lifecycle. OpenClaw says OCE is intended for internal pilot workloads at this stage, ahead of a planned 1.0 release later this year.
OpenClaw targets enterprise deployment barriers
The announcement frames OCE as a response to organizations seeking to deploy persistent agents but wanting common standards for security, safety and governance first. OpenClaw says the lack of those standards has contributed to IT departments banning agent platforms outright. Its stated aim is to let organizations deploy capable agents in enterprise environments while adding controls around how they operate.
OCE is being developed in the open before its 1.0 release. Organizations can self-host it by cloning the project repository and following the getting-started guide. OpenClaw says Docker Compose can be used for local development and Kubernetes for internal deployment. The platform runs on an organization’s own infrastructure and, according to the announcement, will always be free for any organization to use.
Organizations can replace core components
OCE provides shared management and governance while allowing organizations to replace core components, including the agent harness, model and sandbox. The announcement says those components can be third-party products or internal implementations. This vendor-neutral design leaves organizations free to select the tools and environments used to run their agents rather than requiring one fixed combination.
OpenClaw identifies security as the project’s primary focus. Its design combines hard boundaries between trusted and untrusted workloads with sandboxing, reviews performed using large language models and fine-grained permissions. The platform also adds governance and auditability across the agent lifecycle. OpenClaw says it plans to share a reference architecture showing how the protections work together in the coming weeks.
OpenAI and Red Hat are piloting OCE
OpenClaw says OCE began at OpenAI and was later donated to the OpenClaw Foundation. It is now an independent project, further developed in collaboration with Red Hat and NVIDIA. The announcement names OpenAI and Red Hat as organizations piloting the platform internally. It also says OpenAI is already running OpenClaw agents with full access to codebases and plugins.
As an example, the announcement describes Androidclaw, an internal agent used by OpenAI staff to help triage messages in workplace channels. OpenClaw says the agent can use context from tools and services including Git, GitHub and logging systems. Its example involves Androidclaw finding a pull request connected to a broken build and helping fix it.
Our internal enterprise agent Androidclaw has really been well-adopted by our team.
RJ Marsan, Member of the Technical Staff at OpenAI
The example shows the kind of access the project is designed to govern: agents working with internal context and company tools, rather than only responding to isolated prompts. OpenClaw presents OCE as early software for internal trials, with its security approach and deployment controls still being developed ahead of the planned 1.0 release.