Managed Applications Reference | NVIDIA AI Workbench User Guide

This page documents the four types of managed applications in AI Workbench.

For conceptual background, see Applications.
For application fields in spec.yaml, see Project App Information.

Application Types

AI Workbench manages four types of applications.

Each type has different deployment, requirements, and configuration. All application configurations are stored in .project/spec.yaml and versioned in Git.

Type	Description
Web App	An application with a web interface that runs in the project container. AI Workbench proxies it to a local URL for browser access. Requires start/stop commands and a port number. Examples: JupyterLab, TensorBoard, custom dashboards.
Process	A script or background application that runs in the project container without a web interface. Requires start/stop commands but no port configuration. Examples: training scripts, data processing pipelines.
Compose App	A multi-container application defined in a compose file at the repository root. Each service runs in its own container with shared networking. Requires a valid Docker Compose v2 file. See Multi-Container Environments (Docker Compose).
Native App	A system application on your local machine that connects to the project container and repository. Configuration varies by OS and container runtime. Examples: VS Code, Cursor. See Applications and IDEs.

Applications - What applications are and how they work
Walkthrough: Add a Custom Application - Add a custom web application
Project App Information - Application fields in spec.yaml
VS Code - VS Code setup and configuration
Applications and IDEs - Supported IDEs and file editors
Multi-Container Environments (Docker Compose) - Multi-container environment specifications