NVIDIA AI Workbench Error When Running Local

Installed using NIM by default, and “Verify the Installation” showed no issues.

Below are the error log and bundle.zip:

(nvwb:local) workbench@DESKTOP:/mnt/c/Users/user$ nvwb build

failed to create API client

Error:
workbench not available at http://localhost:10001/v1. try deactivating and the re-activating the context. For non-

interactive workflows where the --context and --project flags are used, you must still run nvwb activate first.

ai-workbench-support-bundle.zip (41.4 KB)

When I followed the instructions to open JupyterLab, the webpage couldn’t load and displayed:

“This page isn’t working
localhost didn’t send any data.
ERR_EMPTY_RESPONSE”

However, there were no errors in the cmd.

nvwb start jupyterlab

✅ Container is running
✅ jupyterlab is running

App successfully started and available at:
http://localhost:10000/projects/test/applications/jupyterlab/?token=7ff784a00255615862e9c4e9ab2e56819f27849e2ffee753

hello

to be clear, you used the installer corresponding to this user guide?

then the workbench CLI is functioning but the Desktop App isn’t?

Hello,

Upon reviewing the desktop_app_main.log file we can see that our various WSL calls are failing because of a failure to mount the D drive.

[2025-06-27 13:45:41.009] [info]  (startup-context)                    wsl: Failed to mount D:\, see dmesg for more details.

As the command states you’ll have to use dmesg from within the WSL linux environment to see what’s wrong. You can enter the environment by opening a PowerShell terminal and entering the command wsl -d NVIDIA-Workbench bash.