NVIDIA AI Workbench is a free development environment manager to develop, customize, and prototype AI applications on your GPUs. AI Workbench provides a frictionless experience across PCs, workstations, servers, and cloud for AI, data science, and machine learning (ML) projects. The user experience includes:
- Easy setup on single systems: Click-through install in minutes on Windows, Ubuntu, and macOS, with a one-line install on remote systems.
- Managed experience for decentralized deployment: A free, PaaS/SaaS-type UX in truly hybrid contexts without a centralized, service-based platform.
- Seamless collaboration for experts and beginners: Friendly Git, container, and application management without limiting customization by power users.
- Consistent across users and systems: Migrate workloads and applications across different systems while maintaining functionality and user experience.
- Simplified GPU handling: Handles system dependencies like the NVIDIA Container Toolkit, as well as GPU-enabled container runtime configuration.
- Streamlined multicontainer environments: Create and share multicontainer environments, workloads, and applications with click-through delivery and use.
This post provides details about the January 2025 release of NVIDIA AI Workbench, including the following new integrations and features:
- Local meets cloud with NVIDIA Brev
- NVIDIA AI Blueprint for PDF to podcast in AI Workbench on NVIDIA RTX workstations
- Frictionless collaboration through Git
- New desktop app features
Local meets cloud with NVIDIA Brev
According to feedback received from the recent Dell and NVIDIA HackAI Hackathon, users want easy access to cloud GPUs through AI Workbench. This is now possible, thanks to a collaboration between NVIDIA AI Workbench and NVIDIA Brev, an AI development platform that enables you to run, build, train, and deploy ML models on the cloud.
This AI Workbench release debuts a lightweight integration with Brev for one-click laptop-to-cloud workflows. Users can create a cloud instance in Brev, connect it to AI Workbench, and then launch into the full AI Workbench UX in the cloud.
NVIDIA Brev is a cloud aggregator that enables developers to find and launch GPU instances in minutes, providing access to scalable GPU compute for the best price possible.
Getting started is as easy as one, two, three:
- Create a Brev account
- Install the Brev CLI
- Add your Brev instance as an AI Workbench location using the following commands:
$ brev login$ brev set <org-name>$ nvwb create context --brev-instance-name <my-brev-instance> |
Once bootstrapped, your running cloud instance will automatically appear as an AI Workbench location inside of the AI Workbench desktop application. Simply click into it to work on the instance, or port an existing project from a different location in AI Workbench.
With support for Brev, AI Workbench now enables you to quickly spin up the properly sized instance for your project in the cloud and add the instances as remote locations on your local AI Workbench. This creates an easy local-to-cloud flow for building and testing generative AI applications on any GPU of choice.
For more details, see the AI Workbench and Brev Integration User Guide.
NVIDIA AI Blueprint for PDF to podcast in AI Workbench on RTX workstations
NVIDIA Blueprints are reference workflows that enable developers to get started with agentic AI and generative AI use cases. AI Workbench supports any Docker Compose-based NVIDIA Blueprint with a streamlined, click-through user experience on any suitable GPU enabled system.
AI Workbench is particularly useful for working on your own GPU system, such as an NVIDIA RTX-powered AI workstation. You can easily launch your experience with the multicontainer support feature of AI Workbench, powered by Docker Compose.
A good example of this is the NVIDIA AI Blueprint for PDF to podcast released at CES 2025. This blueprint enables you to build a generative AI application that transforms PDF data—such as training documents, technical research, or documentation—into engaging and personalized audio content.
Built on NVIDIA NIM microservices, this blueprint is flexible and can run securely on a private network, delivering actionable insight without sharing sensitive data. To learn more, see NVIDIA and Partners Launch Agentic AI Blueprints to Automate Work for Every Enterprise.
Running blueprints with AI Workbench projects provides numerous benefits, including:
- Environment and GPU configuration: Clone the PDF to Podcast Blueprint from GitHub and AI Workbench handles the rest with automatic GPU configuration.
- Development integrations: Seamless support for popular development environments such as Jupyter and Visual Studio Code, as well as support for user-configured web applications.
- Containerized and customizable environments: All projects are containerized, isolated, and easily modifiable. Adapt blueprints to suit your specific needs while ensuring consistency and reproducibility.
To get started with this blueprint on AI Workbench, see the Quick Start Guide available through the NVIDIA-AI-Blueprints/pdf-to-podcast GitHub repo.
Frictionless collaboration through Git
The previous release introduced expanded Git functionality in the AI Workbench desktop app and CLI. For example, users can create and manage branches in the Branches view, and see and manage file diffs prior to committing changes. You can use Git through AI Workbench or natively through the terminal.
This release includes additional UI improvements to manage branches and to ensure that file changes are immediately viewable in the desktop app. Users can now create new branches directly from the Branch dropdown menu, in addition to using the Branches view. A user will also be notified when switching branches while running a container.
With an improved UI for branch management, users can develop and experiment independently without affecting the main code, enabling frictionless collaboration across teams and projects.
Learn more about Git in AI Workbench.
New desktop app features
This release introduces new desktop app features that further streamline the user experience, including:
- Filter projects: Users can now filter projects by date or by keywords in the project name or description. This feature is available on the homepage of each location that a user works on.
- Deep link widget: Deep links can now be created on the desktop app in addition to the CLI. Users can generate deep links that, when shared with others, will directly open the project on AI Workbench. Available in the settings page of a project.
- File browser: Edit capability is now possible in the file browser, enabling users to edit files in the project folder directly in the desktop app. This simplifies editing files both locally and remotely, and eliminates the need to start JupyterLab or VS Code to access and edit the files.
Track changes real time: Real-time file watching now updates the view in the desktop app as files change. This fixes a previous lag, where a file might change (name or deletion, for example), but it wouldn’t show immediately in the file browser.
SAGECONNECT TECHNOLOGY GROUP CO., LIMITED
Sageconnect is a reliable provider of IT network system integration and solutions for global enterprise customers.
