OpenShell/docs/index.md

title, description, topics, tags, content

title	description	topics	tags	content
page nav card NVIDIA OpenShell Developer Guide Get Started NVIDIA OpenShell	OpenShell is the safe, private runtime for autonomous AI agents. Run agents in sandboxed environments that protect your data, credentials, and infrastructure.	Generative AI Cybersecurity	AI Agents Sandboxing Security Privacy Inference Routing	type index

NVIDIA OpenShell

NVIDIA OpenShell is the safe, private runtime for autonomous AI agents. It provides sandboxed execution environments that protect your data, credentials, and infrastructure. Agents run with exactly the permissions they need and nothing more, governed by declarative policies that prevent unauthorized file access, data exfiltration, and uncontrolled network activity.

Get Started

Install the CLI and create your first sandbox in two commands.

<style>
.nc-term {
  background: #1a1a2e;
  border-radius: 8px;
  overflow: hidden;
  margin: 1.5em 0;
  box-shadow: 0 4px 16px rgba(0,0,0,0.25);
  font-family: 'SFMono-Regular', Menlo, Monaco, Consolas, 'Liberation Mono', monospace;
  font-size: 0.875em;
  line-height: 1.8;
}
.nc-term-bar {
  background: #252545;
  padding: 10px 14px;
  display: flex;
  gap: 7px;
  align-items: center;
}
.nc-term-dot { width: 12px; height: 12px; border-radius: 50%; }
.nc-term-dot-r { background: #ff5f56; }
.nc-term-dot-y { background: #ffbd2e; }
.nc-term-dot-g { background: #27c93f; }
.nc-term-body { padding: 16px 20px; color: #d4d4d8; }
.nc-term-body .nc-ps { color: #76b900; user-select: none; }
.nc-swap {
  display: inline-grid;
  vertical-align: baseline;
}
.nc-swap > span {
  grid-area: 1 / 1;
  white-space: nowrap;
  opacity: 0;
  animation: nc-cycle 12s ease-in-out infinite;
}
.nc-swap > span:nth-child(2) { animation-delay: 3s; }
.nc-swap > span:nth-child(3) { animation-delay: 6s; }
.nc-swap > span:nth-child(4) { animation-delay: 9s; }
@keyframes nc-cycle {
  0%, 3%     { opacity: 0; }
  5%, 20%    { opacity: 1; }
  25%, 100%  { opacity: 0; }
}
.nc-hl { color: #76b900; font-weight: 600; }
.nc-cursor {
  display: inline-block;
  width: 2px;
  height: 1.1em;
  background: #d4d4d8;
  vertical-align: text-bottom;
  margin-left: 1px;
  animation: nc-blink 1s step-end infinite;
}
@keyframes nc-blink { 50% { opacity: 0; } }
</style>
<div class="nc-term">
  <div class="nc-term-bar">
    <span class="nc-term-dot nc-term-dot-r"></span>
    <span class="nc-term-dot nc-term-dot-y"></span>
    <span class="nc-term-dot nc-term-dot-g"></span>
  </div>
  <div class="nc-term-body">
    <div><span class="nc-ps">$ </span>uv pip install openshell</div>
    <div><span class="nc-ps">$ </span>openshell sandbox create <span class="nc-swap"><span>-- <span class="nc-hl">claude</span></span><span>--from <span class="nc-hl">openclaw</span></span><span>-- <span class="nc-hl">opencode</span></span><span>-- <span class="nc-hl">codex</span></span></span><span class="nc-cursor"></span></div>
  </div>
</div>

Run openshell --help in your terminal to see the full CLI reference, including all commands and flags. You can also clone the OpenShell repository and use the /openshell-cli skill to load the CLI reference into your agent.

Proceed to the Quickstart for more details.

---

Explore

::::{grid} 2 2 3 3 :gutter: 3

:::{grid-item-card} About OpenShell 🔗 about/overview :link-type: doc

Learn about OpenShell and its capabilities.

+++ {bdg-secondary}Concept :::

:::{grid-item-card} Quickstart 🔗 get-started/quickstart :link-type: doc

Install the CLI and create your first sandbox in two commands.

+++ {bdg-secondary}Tutorial :::

:::{grid-item-card} Tutorials 🔗 tutorials/index :link-type: doc

Hands-on walkthroughs from first sandbox to custom policies.

+++ {bdg-secondary}Tutorial :::

:::{grid-item-card} Gateways and Sandboxes 🔗 sandboxes/manage-gateways :link-type: doc

Deploy gateways, create sandboxes, configure policies, providers, and community images for your AI agents.

+++ {bdg-secondary}Concept :::

:::{grid-item-card} Inference Routing 🔗 inference/index :link-type: doc

Keep inference traffic private by routing API calls to local or self-hosted backends.

+++ {bdg-secondary}Concept :::

:::{grid-item-card} Reference 🔗 reference/default-policy :link-type: doc

Policy schema, environment variables, and system architecture.

+++ {bdg-secondary}Reference :::

::::

:hidden:

Home <self>

:caption: About NVIDIA OpenShell
:hidden:

Overview <about/overview>
How It Works <about/architecture>
Supported Agents <about/supported-agents>
Release Notes <about/release-notes>

:caption: Get Started
:hidden:

Quickstart <get-started/quickstart>
tutorials/index

:caption: Gateways and Sandboxes
:hidden:

sandboxes/index
Sandboxes <sandboxes/manage-sandboxes>
Gateways <sandboxes/manage-gateways>
Providers <sandboxes/manage-providers>
Policies <sandboxes/policies>
Community Sandboxes <sandboxes/community-sandboxes>

:caption: Inference Routing
:hidden:

inference/index
inference/configure

:caption: Reference
:hidden:

reference/gateway-auth
reference/default-policy
reference/policy-schema
reference/support-matrix

:caption: Resources
:hidden:

resources/license