On September 29, 2026 the OpenClaw project released OpenClaw Enterprise (OCE), a free, open‑source control plane for managing AI agents at scale. The software is licensed under the MIT License and can be installed on a company’s own Kubernetes cluster or run locally with Docker Compose. OpenAI, Red Hat and Nvidia are listed as backing partners, and the code is hosted on GitHub (github.com/openclaw/openclaw‑enterprise). OCE adds multi‑tenant administration, fine‑grained permissions, sandboxing and audit logs to the original OpenClaw framework, which started as a simple “AI harness” that could read a user’s email and reply automatically.
What does OCE actually do?
Think of OCE as “Kubernetes for agents.” It provides the OpenClaw Control Plane (OCC) that handles agent deployment, lifecycle management, and policy enforcement. Companies can swap in their own large language models (LLMs), storage back‑ends, or sandbox implementations, keeping the underlying infrastructure under their control. The platform includes a Kubernetes‑only profile for production use and a Docker Compose preview for developers. While the software itself is free, organizations still need to pay for the compute, storage and any third‑party models they run.
Legal and security backdrop
The Register reported that OpenClaw faces a lawsuit aimed at preventing the spread of the tool in businesses that have banned it for security reasons. Analysts have called the original OpenClaw project an “unacceptable cybersecurity risk” and warned about weak default configurations. OCE is positioned as a response to those concerns, adding hard security boundaries, RBAC (role‑based access control) and audit trails so IT teams can safely let agents touch production systems.
Why it matters
For a regular company, OCE means they can try out persistent AI assistants without sending data to a third‑party SaaS. The built‑in security and audit features aim to satisfy IT policies that have so far forced many firms to ban such agents.
