Welcome to Brev

NVIDIA’s Brev platform is your starting point for active AI development and experimentation. Brev eliminates barriers to AI/ML development by providing access to a broad range of GPUs. We aggregate multiple cloud providers to find optimal GPU configurations. Developers simply choose their requirements and begin coding without extensive infrastructure management.

Core Features and Capabilities

Brev offers the following key features and functionality:

• GPU-backed Sandboxes: Virtual machines with NVIDIA GPU access, preconfigured with Python, CUDA, Docker, and Jupyter Notebooks for rapid testing and prototyping.

• Frictionless Development through Launchables: Take your GPU-backed Sandbox further with fully optimized environments that deploy with a single click. Customize them with specific GPU resources, Docker images, setup scripts, and project files. Then, easily share across your teams for consistent development environments and faster onboarding.

• CLI and Multi-IDE Support: Interact with your Brev environments using Brev CLI commands, and integrate with your existing Cursor, Windsurf, code-server, or tmux.

• Serverless Deployments: Deploy select NVIDIA NIM microservices as auto-scaling serverless functions for development against model-specific APIs.

To start, click here to create your first GPU instance and experience fully-configured NVIDIA GPU environments.