Agents
zeroboot
Sub-millisecond VM sandboxes for AI agents via copy-on-write forking.
★ 2.5k starsApache-2.0
View guide →NVIDIA's reference stack for running AI agents more securely inside OpenShell sandboxes
NemoClaw is an open-source ai agents project by NVIDIA (NVIDIA/NemoClaw on GitHub). It has 22,686 GitHub stars and 3,131 forks, is released under the Apache-2.0 license, and was last updated on GitHub on 8 October 2026.
Confirm you have a supported host - the docs target DGX systems or Windows Subsystem for Linux (WSL).
Clone the repository and open the linked quickstart for the agent you want to run.
Run the terminal installer, or let your coding agent install it with you using the official starter prompt.
Accept the express install preset or choose your agent, sandbox name, inference provider and model interactively.
Runs agents inside NVIDIA OpenShell so their work is contained rather than executed directly on the host.
Supports OpenClaw (the express-install default), Hermes and LangChain Deep Agents Code with dedicated quickstarts.
Connects sandboxed agents to managed inference providers and models chosen at install time.
A preset-driven installer configures recommended settings per platform, with an interactive mode when you want control.
An official starter prompt lets Cursor, Claude Code, Codex or Copilot install NemoClaw with you, asking one question at a time.
Ships with docs covering prerequisites, per-agent quickstarts and user guides.
Get the source code from GitHub and enter the project directory.
git clone https://github.com/NVIDIA/NemoClaw.git
cd NemoClawThe README routes installation through the official docs and terminal installer rather than a single package command.
# See the prerequisites and quickstart for your agent in the docs linked from the READMEExpress install sets up OpenClaw by default; other guides cover Hermes and LangChain Deep Agents.
# Pick a quickstart (OpenClaw / Hermes / Deep Agents) from the docs indexIt is maintained by NVIDIA as an open-source reference stack on GitHub.
Yes - NemoClaw is Apache-2.0 licensed and free to use under that license.
The software is free, but inference comes from a provider you choose during setup - that provider may require its own account or API key, and local managed inference depends on your hardware.
Compared with running OpenClaw or a similar agent directly on your machine, NemoClaw adds OpenShell sandboxing and managed inference as the default posture, trading some simplicity for isolation.
Check the prerequisites in the docs, clone the repo, and follow the quickstart for OpenClaw, Hermes or Deep Agents.
Sub-millisecond VM sandboxes for AI agents via copy-on-write forking.
NVIDIA's object-oriented, Pythonic approach to building AI agents.
ByteDance's open-source multimodal AI agent stack: GUI agents that see and operate real desktops
Secure sandboxes where AI agents run real code.
The open runtime for AI agents — a Rust platform for running, governing and routing agents with TokenOps.
NVIDIA's toolkit for connecting, profiling and optimizing teams of AI agents — across any framework
The most-starred other AI Agents projects in the LibreAI catalogue — the realistic shortlist if NemoClaw is not the right fit.
A personal AI that really does things — on any OS, with your data staying yours.
An agentic skills framework & software development methodology that works.
Agent harness optimization system with skills, memory and security for Claude Code, Codex and more
The agent that grows with you — a Python framework for adaptive AI agents.
DeepSeek's plugin-based agent harness where everything is a plugin.
Free and open — grab the source, star the repo, and start creating.