NVIDIA NemoClaw#

GitHub License

NemoClaw is the OpenClaw plugin for NVIDIA OpenShell. It runs OpenClaw inside a sandboxed environment with NVIDIA inference, such as Nemotron 3 Super 120B through build.nvidia.com or local vLLM. The sandbox enforces strict network policies and operator-controlled egress approval.

Get Started#

Install the CLI and launch a sandboxed OpenClaw instance in a few commands.

$ git clone https://github.com/NVIDIA/NemoClaw.git
$ cd NemoClaw
$ ./install.sh

Run nemoclaw --help in your terminal to view the full CLI reference. You can also clone the NemoClaw repository to explore the plugin source and blueprint.

Proceed to the Quickstart for step-by-step instructions.


Explore#

About NemoClaw

Learn what NemoClaw does and how it integrates OpenClaw with OpenShell.

Overview
Quickstart

Install the CLI, configure inference, and launch your first sandboxed agent.

Quickstart
Commands

CLI commands for launching, connecting, monitoring, and managing sandboxes.

Commands
Inference Profiles

Switch between NVIDIA cloud, local NIM, and vLLM inference backends.

Inference Profiles
How It Works

High-level overview of the plugin, blueprint, sandbox, and inference routing.

How NemoClaw Works
Architecture

Plugin structure, blueprint system, and sandbox lifecycle.

Architecture
Network Policies

Egress control, operator approval flow, and policy configuration.

Network Policies
How-To Guides

Task-oriented guides for inference, deployment, and policy management.

Switch Inference Providers at Runtime