> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.nvidia.com/ai-workbench/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.nvidia.com/ai-workbench/_mcp/server.

> 'AI Workbench Walkthrough: Customize Your Environment (CLI)'

# Walkthrough: Customize Your Environment (CLI)

This section of the documentation is a walkthrough that shows you how to customize your container environment by adding a package to your [NVIDIA AI Workbench](/ai-workbench/overview/introduction) project.
For the full list of quickstarts and walkthroughs, see [Quickstarts and Walkthroughs](/ai-workbench/quickstart/overview).

The base container that you choose when you create your AI Workbench project might not have all the features that you need.
You can customize your project by installing additional packages, customizing mounts, and defining new environment variables.

In this walkthrough, you use the [AI Workbench CLI](/ai-workbench/reference/user-interface/cli) to perform the following tasks:

- [Start AI Workbench and Open a Project](/ai-workbench/quickstart/quickstart-environment-cli#start-ai-workbench-and-open-a-project)
- [Customize Your Project by Installing a Package](/ai-workbench/quickstart/quickstart-environment-cli#customize-your-project-by-installing-a-package)
- [Rebuild Your Project Environment](/ai-workbench/quickstart/quickstart-environment-cli#rebuild-your-project-environment)
- [Test by Using Your New Package in Code](/ai-workbench/quickstart/quickstart-environment-cli#test-by-using-your-new-package-in-code)

<Note>

To create your own fully custom container, see [Advanced Walkthrough: Use Your Own Container](/ai-workbench/quickstart/quickstart-byoc) instead.

</Note>

## Prerequisites

Before you complete the steps in this walkthrough, you need the following:

- NVIDIA AI Workbench is installed on your local system.
- A basic test project. If you don't have one, see [Basic Quickstart (CLI)](/ai-workbench/quickstart/quickstart-cli).

## Start AI Workbench and Open a Project

1. Open a command shell by doing one of the following:

   - **On macOS or Ubuntu** — Open a command shell.

   - **On Windows** — Open the WSL app by searching for WSL in your app list. When the WSL app opens, you should see a command prompt similar to the following:

     ```shell
     workbench@computer-name:~$
     ```

     <Tip>

     You can also use Terminal, Command Prompt, or Windows PowerShell, and access WSL by using the command `wsl -d NVIDIA-Workbench`.

     </Tip>

2. Check the available locations (contexts) on your computer by running the following command.

   ```shell
   nvwb list contexts
   ```

   If you haven't added any remote locations yet, the only available location is the local computer. The output should look similar to the following.

   ```shell
     NAME  | DESCRIPTION | HOSTNAME  |  STATUS
   --------|-------------|-----------|---------------
     local | My Computer | localhost | Not Running
   ```

3. Run the following command to start the local location.

   ```shell
   nvwb activate local
   ```

   AI Workbench starts and connects to your computer, and `(nvwb:local)` now appears at the beginning of your command prompt.

4. Run the following command to see the names and statuses of your projects.

   ```shell
   nvwb list projects
   ```

5. Run the following command to open your project. Change `<your project name>` to the name of your project.

   ```shell
   nvwb open <your project name>
   ```

   AI Workbench opens your project, changes to the project directory, and `(nvwb:local/<your project name>)` now appears at the beginning of your command prompt.

## Customize Your Project by Installing a Package

Use the following procedure to add a package to your AI Workbench project.

1. Run the following command to open JupyterLab.

   JupyterLab opens in your default browser ready for you to start working.

   ```shell
   nvwb start jupyterlab
   ```

   JupyterLab opens in your default browser ready for you to start working.

2. In JupyterLab, in the file browser, double-click *requirements.txt* to open it.

   The file opens and you should see an entry for `jupyterlab`.

3. Move to the first empty line in the file, type `numpy`, and then save the file.

4. In JupyterLab, choose **File**, and then click **Shut Down**. Confirm shut down and close the browser tab.

## Rebuild Your Project Environment

Whenever you change your project environment, you must rebuild the project before you can access the changes. Use the following procedure to rebuild your environment.

1. In your command shell, check the status of your project by running the following command.

   ```shell
   nvwb status
   ```

   You should see output similar to the following.

   ```shell
   Name: test-cli-proj (/home/workbench/nvidia-workbench/test-cli-proj)
   Description: A hello world test project
   Environment: Full Build Required
   Container: Running
   Applications:
     jupyterlab: Stopped
   Remote Repository: https://github.com/<user_name>/test-cli-proj.git (Up-to-date)
   Local Repository: 1 Modified
   ```

   Line 8 indicates that you changed your project by editing the requirements.txt file. Line 3 indicates that the changes you made require you to rebuild your project.

2. Run the following command to stop your container environment. If the container is already stopped you see a message that the container is not running. Continue to the next step.

   ```shell
   nvwb stop --container
   ```

3. Run the following command to rebuild your project.

   ```shell
   nvwb build
   ```

   The project builds. Wait until you see **Container build complete** and then go to the next section to test your new package.

4. Run the following command to start your container environment.

   ```shell
   nvwb start --container
   ```

## Test by Using Your New Package in Code

After you add the package to your project and rebuild the container environment, you can use the new package in your code.

1. To open JupyterLab, run the following command.

   JupyterLab opens in your default browser ready for you to start working.

   ```shell
   nvwb start jupyterlab
   ```

2. In JupyterLab, navigate to the *code* folder. If your project already has a notebook, click to open it. Otherwise; create a new notebook by choosing Python 3 (ipykernel).

3. In your notebook, enter the following code and then click run.

   ```python
   import numpy as np
   ```

4. Enter the following code and then click run.

   ```python
   print(np.e)
   ```

   You should see the following output which is the value of `e`.

   ```python
   2.718281828459045
   ```

5. Save your notebook.

6. In JupyterLab, choose **File**, and then click **Shut Down**. Confirm shut down and close the browser tab.

7. In your command shell, run the following command to stop your project environment.

   ```shell
   nvwb stop --container
   ```

## Related Topics

- [Walkthrough: Customize Your Environment](/ai-workbench/quickstart/quickstart-environment)
- [Walkthrough: Add a Custom Application](/ai-workbench/quickstart/quickstart-custom-app)
- [Advanced Walkthrough: Hybrid RAG](/ai-workbench/quickstart/quickstart-hybrid-rag)
- Explore [NVIDIA AI Workbench Example Projects](/ai-workbench/quickstart/example-projects)
- Explore the [NVIDIA Developer Forum](https://forums.developer.nvidia.com/c/ai-data-science/nvidia-ai-workbench/671)

**Learn the Concepts**

- [How to Work with AI Workbench Projects](/ai-workbench/how-to/projects/overview)
- [Add a Brev Instance](/ai-workbench/how-to/locations/add-brev)
- [Integrations and Credentials](/ai-workbench/reference/faq#integrations-and-credentials)
- [Applications](/ai-workbench/concepts/application-concept)