Skip to main content
OpenFree logoOpenFree
AdOpenFree logoPromote your productReach more potential users and drive product growth and revenue.Advertise
Favicon of NVIDIA NemoClaw

NVIDIA NemoClaw

Free Listing

Open source reference stack from NVIDIA for running supported and qualification-only AI agents more safely inside OpenShell sandboxes, with guided onboarding, managed inference, network policy, integrations, and lifecycle operations.

Visit NVIDIA/NemoClaw

Overview

NVIDIA NemoClaw is an open source reference stack for running supported and qualified candidate AI agents more safely inside NVIDIA OpenShell sandboxes. It installs OpenShell as part of NVIDIA Agent Toolkit, applies the selected agent integration layer and versioned blueprint, and provides guided onboarding, managed inference, network policy, managed integrations, and lifecycle operations.

Key Features

  • Guided onboarding and an onboard wizard for setup.
  • Managed inference with credential custody provided by OpenShell.
  • Network policy and managed integrations.
  • Lifecycle operations for sandboxed agents.
  • Separation of qualification-only candidate agents from supported agent runtimes.
  • Versioned blueprints and selected agent integration layers.

Use Cases

  • Running supported and qualified candidate AI agents inside NVIDIA OpenShell sandboxes.
  • Onboarding OpenClaw, Hermes, or LangChain Deep Agents Code as sandboxed agents.
  • Configuring messaging channels during initial onboarding for OpenClaw or Hermes.

Getting Started

  • Install NemoClaw and run the onboard wizard to get started.
  • Use the terminal installer for the default installation process; it starts the maintained last-known-good installer and onboard wizard.
  • By default, NemoClaw installs the OpenClaw agent; separate quickstarts cover OpenClaw, Hermes, LangChain Deep Agents Code, and Pi.

Deployment & Requirements

  • Run agents inside NVIDIA OpenShell sandboxes; NemoClaw installs OpenShell, part of NVIDIA Agent Toolkit.
  • NemoClaw applies the selected agent integration layer and a versioned blueprint.
  • OpenShell provides credential custody for managed inference and host-configured integration credentials.

Before You Adopt

  • License: Apache-2.0. Review its terms before using, modifying, or distributing the project.
  • Qualification-only candidate agent runtimes are kept separate from supported runtimes.
  • Installer and supported CLI operations attempt to send operation telemetry to NVIDIA’s fixed Production receiver by default; telemetry can be disabled.

Categories

Agent SDKs & Runtimes

Tags

ai-agentclinvidiasandboxsecurityterminal

Comments

Sign in to leave a comment.

More like NVIDIA NemoClaw

OpenFree logo

vork

Free ListingStars: 1K

Headless 3D and VFX engine built for agents

Agent SDKs & Runtimes

vork is a headless 3D and VFX engine for agents. Agents write scenes as code, render and inspect results, then revise. Written in Rust and Slang, it renders on GPUs through Vulkan and Metal.

Favicon of AX

AX

Free ListingStars: 14.4K

Declarative orchestration for agentic workloads at scale

Agent SDKs & Runtimes

AX is Google's open agentic orchestration runtime that sandboxes tasks, configures workspaces, enforces network policies, and manages models so teams can run large numbers of agent workloads per cluster.

Favicon of Moli

Moli

Free ListingStars: 20.6K

Open-source headless browser for AI agents

Agent SDKs & Runtimes

Moli is a production-ready, open-source headless browser for AI agents, built in Rust. It fetches and extracts web pages, searches the web, and automates browser tasks via CLI, CDP, and WebDriver with a structure-first design.