Build a Claw: NVIDIA NemoClaw on DGX Spark | Nemotron Labs

Want to build a long‑running AI agent on Nemotron open models using OpenClaw? Recently announced at NVIDIA GTC San Jose, NemoClaw (in alpha) is an open source reference stack that launches autonomous agents built with OpenClaw inside OpenShell’s sandboxed environment—so with one command you can operate always‑on, self‑evolving assistants more safely. OpenShell (also in alpha) is the runtime and policy layer that adds enterprise‑grade privacy and security guardrails. In this Nemotron Labs livestream, you can: Watch a step‑by‑step install of NemoClaw on DGX Spark and see what it takes to get a claw running end‑to‑end. Learn how NemoClaw uses open models like NVIDIA Nemotron together with the NVIDIA OpenShell runtime to provide a safer environment for executing claws. Get a clear explanation of how NemoClaw, OpenShell, and OpenClaw fit together so you know where each piece runs and how to start experimenting confidently. Join the stream, bring your installation and deployment questions, and come ready with a DGX Spark if you have one so you can follow along and run through the installation live alongside us. __________ *Developer Resources:* _Nemotron 3 Super — DGX Spark Deployment Guide → https://github.com/NVIDIA-NeMo/Nemotron/tree/main/usage-cookbook/Nemotron-3-Super/SparkDeploymentGuide_ _NeMoClaw OpenShell Install Docs → https://github.com/NVIDIA/NemoClaw/blob/main/spark-install.md_ _OpenShell Github → https://github.com/NVIDIA/OpenShell_ _*NeMoClaw on Build → https://nvda.ws/4c05SmK*_ NeMoClaw Discord → https://discord.com/channels/1019361803752456192/1482072289511211200 _*NeMoClaw & OpenShell blog → https://nvda.ws/3N6I7RV*_ PinchBench with Nemotron 3 Super → https://pinchbench.com