Nvidia Introduces NemoClaw an AI Stack to Make OpenClaw Agents More Secure

On Monday Nvidia introduced NemoClaw, an artificial intelligence (AI)-powered stack for the OpenCla agents, on its own initiative. In the keynote session, it announced at the company’s annual GPU Technology Conference (GTC) that its CEO was a “regular processor” in their business. NemoClaw essentially adds privacy and security guardrails to OpenC Law’s AI agents, which are deployable and scalable for business applications. The tech giant, based in Santa Clara, California, noted that the AI stack can be installed with one command and is platform and agent agnostic. Developers and enterprises are now able to use NemoClaw as a preview of NemeC law.

Nvidia Introduces NemoClaw

tech giant, introducing and detailing NemoClaw in an article on the newsroom that was written by him. A good one, since it was specifically developed for OpenClaw (although it can also work with other agents) is the new offering. After it became popular with developers for its general-purpose agents, called claws, the agentic automation company was recently acquired by OpenAI.

Like most AI agents, which are built for specific use cases and have limited access and knowledge of external tools, OpenClaw’s agents can perform a wide variety of tasks in vogue environments; they employ many different types of tools. Its talent, however, is also its biggest vice-versa.

Since OpenClaw can access the Internet, many databases and the device it is on, it becomes a target for cyber criminals. The lack of standardised security and privacy protections in the open-source tool has also led to a low adoption from businesses. This is the problem Nvidia hopes to solve with NemoClaw.

The software used by NemoClaw, which uses Nvidia’s Agent Toolkit to optimise OpenCla for use in businesses and specific developer environments. In its statement, the company says it installs the open-source OpenShell runtime with one command “I use this service to do so… This runstime also creates an isolated sphere and open models, through which security and privacy guardrails can be added to autonomous agents via the runtime.

According to Nvidia, NemoClaw can use any coding agent to tap into AI models such as the recently released Nemosotron 3 Super (and run locally on users’ own dedicated system). The use of a privacy router allows users to access the cloud’s frontier models by it, which is used in its own work. Now it is available in preview and can be downloaded from the company’s website or GitHub listing. It also allows businesses to deploy, and run AI agents via large-scale cloud service providers.