OpenClaw launches free enterprise control plane for AI agents

Managing a single AI agent is one thing. Managing dozens of them across departments, teams, and security boundaries in a large organization is a completely different beast. OpenClaw thinks it has the answer.

The OpenClaw Foundation launched OpenClaw Enterprise (OCE) on September 29, a free, open-source control plane purpose-built for running persistent AI agents in multi-user and sensitive enterprise environments. The platform adds multi-tenancy, auditability, governance controls, and enhanced security boundaries on top of the existing OpenClaw agent framework, essentially turning what started as a personal AI tool into something an enterprise IT team could actually deploy without losing sleep.

From OpenAI side project to enterprise foundation

OCE’s origin story has an unusual arc. The project’s development began inside OpenAI before being donated to the OpenClaw Foundation, a move that gave it independence from any single vendor while preserving the engineering momentum behind it. Peter Steinberger originally created OpenClaw as an autonomous AI agent framework in late 2025, and the tool has evolved rapidly since.

The foundation has lined up some heavyweight collaborators. Red Hat and NVIDIA are both actively involved, and internal pilot programs are already running at Red Hat and OpenAI.

OpenClaw 2.0 dropped on August 30, roughly a month before OCE’s announcement, introducing collaborative enhancements and security protocols designed to transition the framework from solo developer usage to shared enterprise environments. OCE builds directly on that foundation, adding the organizational layer that companies with compliance requirements and multi-team structures actually need.

What OCE actually does

Multi-tenancy is the headliner feature. In practical terms, this means different teams or business units within the same organization can run their own AI agents without stepping on each other’s toes or accidentally accessing data they shouldn’t. The governance and auditability features sit alongside this, giving administrators visibility into what agents are doing and the ability to enforce policies across the board.

On the deployment side, OCE offers two paths. Teams can spin it up locally using docker-compose for development and testing, or deploy it on Kubernetes for production workloads.

The platform is also model-agnostic, meaning it doesn’t lock organizations into any particular AI model provider.

Kevin Lin, involved with the project, noted that enterprises want the flexibility of open source combined with the governance frameworks they’re accustomed to in traditional software.

The open-source bet

Perhaps the most notable aspect of OCE is what it doesn’t include: a price tag. There are no paid tiers, no hosted services, no premium features locked behind a subscription. The entire platform is self-hosted on user infrastructure.

Where this fits in the enterprise AI stack

Red Hat’s involvement is particularly telling. The company built its entire business on packaging and supporting open-source software for enterprises. Its willingness to pilot OCE internally suggests the platform has passed at least a baseline threshold of enterprise readiness, even before the official 1.0 release.

Version 1.0 is expected in the near future, and OCE is currently in the internal pilot phase at Red Hat and OpenAI before that full release.

Disclosure: This article was edited by Diego Almada Lopez. For more information on how we create and review content, see our Editorial Policy.