User Guide (Latest)

In NVIDIA AI Workbench, a computer that has AI Workbench installed is named a location. A location can be CPU-only or it can have a GPU. The following are examples of AI Workbench locations:

  • Laptop

  • Workstation

  • Virtual Machine

  • Cloud Instance

Location Types

  • Local AI Workbench: This is the first and main place you install AI Workbench. It provides the user interface for working locally or with a Remote AI Workbench Location. You only need one Local instance.

  • Remote AI Workbench: This is a system that you have SSH access to. You install AI Workbench on it and connect it to your Local AI Workbench. You can have multiple Remote instances.


The main Location is your system. It has a view of the different Locations it is connected to.

Operating Systems

  • Local: Windows 11 (build 22000 or higher), Windows 10 (build 19044 or higher), Ubuntu 22.04, or macOS (12 or higher)

  • Remote: Ubuntu 22.04

One UI, Many Locations

  • You can add and delete Remote Locations using the Desktop App or CLI on your laptop.

  • Selecting a Location will connect you to it and you can work on the Projects in that Location.

  • Locations don’t need a GPU. They just need AI Workbench installed.


Adding a Remote Location to your Local AI Workbench gives you easy access to more scalable resources for development and compute.

Separate File Systems and Different Resources

  • You can use different GPU-enabled systems within the same UI/UX driven from your laptop.

  • You can create and clone Projects in one Location and easily move them to another location with only minor changes to reflect different system resources like GPUs.

  • You can scale up and scale down


Previous NVIDIA AI Workbench Example Projects
Next Workbench Projects
© Copyright © 2024, NVIDIA Corporation. Last updated on Jun 10, 2024.