NVIDIA OpenShell
https://github.com/NVIDIA/OpenShell
NVIDIA OpenShell is an open-source runtime for running AI agents in isolated sandboxes governed by filesystem, process, and network policies. Developers and teams use it for coding agents and other agent workloads on their own machines or Kubernetes infrastructure.
| Pricing model | Open-source, self-hosted; users provide the compute infrastructure |
|---|---|
| Pricing The headline compute price as the provider states it. Sorting ranks prices stated per vCPU over time, converted to hourly. Other pricing models sort after, unranked. | No OpenShell compute rate published; compute is user-provided |
| Free while idle Whether a stopped, paused, or sleeping sandbox costs nothing. | – |
| Elastic Resources flex with what the sandbox actually uses, and the bill follows: yes when the provider bills on active or observed use, no when capacity is reserved or allocated and billed while it runs. | – |
| Memory snapshots Whether RAM state survives a pause and resume, not just disk. | ✕ |
| Wake on request Whether a stopped sandbox wakes automatically on inbound traffic. | ✕ |
| Start / resume Typical time from create or resume to running. Sorting ranks stated times, converted to seconds; claims with no figure sort after, unranked. | – |
| Max runtime The longest a sandbox may run. | – |
| Isolation The isolation technology between sandboxes. | Configurable Docker or Podman containers, Kubernetes pods, or VMs, with Landlock, seccomp, and network-namespace controls on the agent process |
| GPUs Whether GPU instances are available for sandboxes. | ✓ |
| Docker Whether Docker containers can run inside the sandbox. | ✕ |
| SDKs | Python, TypeScript, Go, Rust |
✓ yes · ✕ no · – no cited public fact.
GPU allocation depends on the runtime and host setup: Docker and Podman require NVIDIA CDI configuration, Kubernetes sets the nvidia.com/gpu limit, and VM gateways accept only one GPU. MicroVMs require host virtualization (Apple Hypervisor on macOS or KVM on Linux).
Sources
- https://docs.nvidia.com/openshell/about/overview.md
- https://docs.nvidia.com/openshell/about/how-it-works.md
- https://github.com/NVIDIA/OpenShell
- https://github.com/NVIDIA/OpenShell/blob/main/architecture/sandbox.md
- https://docs.nvidia.com/openshell/latest/how-it-works/sandboxes/overview.md
- https://docs.nvidia.com/openshell/latest/how-it-works/sandboxes/runtimes
- https://docs.nvidia.com/openshell/latest/how-it-works/gateways/overview
- https://github.com/NVIDIA/OpenShell/issues/2652
- https://github.com/NVIDIA/OpenShell/issues/113
- https://github.com/NVIDIA/OpenShell/blob/main/architecture/compute-runtimes.md
Last verified 2026-10-04. Corrections land through the update log.