# NVIDIA/NemoClaw

Run agents like Hermes, LangChain Deep Agents, and OpenClaw more securely inside NVIDIA OpenShell with managed inference

Repository: https://github.com/NVIDIA/NemoClaw
Canonical: https://ross.abutalabs.com/products/nemoclaw
Homepage: https://docs.nvidia.com/nemoclaw/latest/
Language: TypeScript
License: Apache-2.0
License Family: permissive
Topics: ai-agents, nvidia, openclaw, openshell, sandboxing, typescript, hermes
Last push: 2026-08-27T00:28:07+00:00

## Health v2 (maintenance only)
Score: 59/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 99, release rhythm 35, longevity 12
- inputs: {"age_days": 171, "days_push": 7, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, young
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 22287, forks 3052 (observed 2026-08-28T04:11:32.522650+00:00)

## What it is
NVIDIA NemoClaw is an open-source reference stack for running AI coding agents such as OpenClaw, Hermes, and LangChain Deep Agents Code securely inside NVIDIA OpenShell sandboxes. It provides a CLI with agent-specific aliases, managed inference routing, network policy enforcement, snapshots, credential management, and lifecycle operations for sandboxed agent deployments.

## Use cases
- run OpenClaw or Hermes agents in a secure sandbox
- manage LLM inference routing for AI agents
- enforce network policies on autonomous coding agents
- set up sandboxed AI agents on DGX or WSL hosts
- manage credentials and secrets for AI agents
- monitor and control what sandboxed agents can access
- deploy AI agents to a headless server securely

## When to choose
- you want to run OpenClaw, Hermes, or LangChain Deep Agents with strong sandboxing and network controls
- you need managed, centralized inference routing for multiple agents
- you're on NVIDIA DGX hardware or WSL and want a guided installer
- you need enterprise-style security controls, credential rotation, and monitoring for agent workloads

## When to avoid
- you only need a simple agent framework without sandboxing overhead
- you don't use any of the supported agents (OpenClaw, Hermes, LangChain Deep Agents)
- you need a lightweight library to embed in your own app rather than an ops stack
- you're not prepared to run Docker-based sandbox infrastructure

## Facets
- artifact type: framework
- maturity: active
- function: agent-framework, security, llm-inference, mcp, cli, monitoring, networking
- domain: security, developer-tools, large-language-models, self-hosted
- platform: windows, cli, self-hosted
- tags: sandboxing, openclaw, openshell, hermes, langchain-deep-agents, network-policy, managed-inference, nvidia, dgx, wsl, ai-agents, linux, docker, typescript

## Member repositories
- NVIDIA/NemoClaw (main) score 59

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:11:32.522650+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T16:57:22.317992+00:00, confidence not recorded.
  - readme: https://github.com/NVIDIA/NemoClaw (fetched 2026-08-28T04:11:32.522650+00:00, sha 6d15f3c7394a)
  - homepage: https://docs.nvidia.com/nemoclaw/latest/ (fetched 2026-08-29T07:56:10.423758+00:00, sha 9d6b9a0f0a6f)
  - site_page: https://docs.nvidia.com/nemoclaw/user-guide/openclaw/inference/about-inference-routing (fetched 2026-08-29T07:56:10.437818+00:00, sha d274b1659465)
  - site_page: https://docs.nvidia.com/nemoclaw/user-guide/openclaw/manage-sandboxes/install-openclaw-plugins (fetched 2026-08-29T07:56:10.439700+00:00, sha de20b7b05d99)
  - site_page: https://docs.nvidia.com/nemoclaw/user-guide/openclaw/resources/agent-skills (fetched 2026-08-29T07:56:10.441850+00:00, sha 891b3a6ac983)
  - site_page: https://docs.nvidia.com/nemoclaw/user-guide/openclaw/resources/engineer-agentic-documentation (fetched 2026-08-29T07:56:10.444083+00:00, sha 11cb5c91c865)
  - site_page: https://docs.nvidia.com/nemoclaw/user-guide/openclaw/about/overview (fetched 2026-08-29T07:56:10.427551+00:00, sha d74abbf74569)
  - site_page: https://docs.nvidia.com/nemoclaw/user-guide/openclaw/about/how-it-works (fetched 2026-08-29T07:56:10.429729+00:00, sha 2e46843c65a9)
  - site_page: https://docs.nvidia.com/nemoclaw/user-guide/openclaw/about/ecosystem (fetched 2026-08-29T07:56:10.431764+00:00, sha 598e38215452)
  - site_page: https://docs.nvidia.com/nemoclaw/user-guide/openclaw/get-started/quickstart (fetched 2026-08-29T07:56:10.434244+00:00, sha 890adea5a060)
- Data as of 2026-08-30T08:39:29.467469+00:00.