About vCluster
vCluster creates virtual Kubernetes clusters that run inside a shared host cluster. It is used to give each tenant or team its own isolated environment, including environments for AI workloads.
What it does
Each tenant sees only its own cluster, so it does not see the host's internals or other tenants. The project is open source under the Apache-2.0 license. The company also offers a wider platform for managing GPU infrastructure, bare metal machines and many cluster types. Operators can sell clusters as a managed product, either through a built-in interface or through their own portal using APIs.
Key features
- Virtual Kubernetes clusters for each tenant
- Isolated clusters for Slurm and Ray workloads
- Cluster templates and capacity management
- Node provisioning and autohealing
- Web interface and APIs for self-service
Who it's for
Platform teams at AI cloud providers, enterprises that run internal AI factories, and companies that want to offer Kubernetes to several teams without giving each one a separate physical cluster.
Pricing and good to know
The open source project is free to use under Apache-2.0. The commercial platform is priced separately, so check the pricing page for current plans. Test the open source version on a small cluster before planning production use, and review how the host is configured, since isolation depends on that setup.
Highlights
- Create a virtual cluster for each tenant
- Run isolated clusters for Slurm and Ray
- Manage cluster templates and capacity
- Provision and heal nodes automatically
