### Install Brev CLI on Windows (WSL)

Source: https://docs.nvidia.com/brev/cli/getting-started.md

Install Ubuntu via WSL and run the installation script within the Linux environment.

```powershell
# Install WSL with Ubuntu (run in PowerShell as admin)
wsl --install -d Ubuntu-22.04

# Restart your computer, then open Ubuntu and run:
bash -c "$(curl -fsSL https://raw.githubusercontent.com/brevdev/brev-cli/main/bin/install-latest.sh)"
```

--------------------------------

### ML Training Setup Workflow

Source: https://docs.nvidia.com/brev/cli/composable-workflows.md

Chain search, creation, setup scripts, and IDE opening into a single command pipeline.

```bash
# Find a cheap A100, create with setup script, and open
brev search --gpu-name A100 --sort price \
  | brev create training-box \
  | brev exec @setup.sh \
  | brev open cursor
```

--------------------------------

### Install Brev CLI on Linux

Source: https://docs.nvidia.com/brev/cli/getting-started.md

Download and execute the latest installation script.

```bash
# Download and run the installer
bash -c "$(curl -fsSL https://raw.githubusercontent.com/brevdev/brev-cli/main/bin/install-latest.sh)"

# Verify installation
brev --version
```

--------------------------------

### Use setup scripts

Source: https://docs.nvidia.com/brev/cli/instance-management.md

Automate environment configuration by running scripts after instance startup.

```bash
# Run a setup script from URL
brev start https://github.com/user/repo.git --setup-script https://raw.githubusercontent.com/user/repo/main/setup.sh

# Run a script from within the repo
brev start https://github.com/user/repo.git --setup-repo . --setup-path ./scripts/setup.sh
```

--------------------------------

### Start or create an instance

Source: https://docs.nvidia.com/brev/cli/instance-management.md

Start a stopped instance or create a new one from a Git URL or local path.

```bash
brev start my-instance
```

--------------------------------

### Install and Test ML Framework in Container

Source: https://docs.nvidia.com/brev/troubleshooting/instances-gpus/gpu-detection-pytorch.md

Install Unsloth and verify the import inside the container environment.

```bash
pip install -U unsloth unsloth-zoo
python - <<'PY'
from unsloth import FastLanguageModel
print("Unsloth import OK")
PY
```

--------------------------------

### Install via Homebrew

Source: https://docs.nvidia.com/brev/troubleshooting/cli-installation/cli-installation-linux.md

Alternative installation method using the Homebrew package manager.

```bash
brew install brevdev/homebrew-brev/brev
```

--------------------------------

### Manual installation script execution

Source: https://docs.nvidia.com/brev/troubleshooting/cli-installation/cli-installation-linux.md

Steps to download, set permissions, and execute the installation script manually to capture potential errors.

```bash
curl -fsSL https://raw.githubusercontent.com/brevdev/brev-cli/main/bin/install-latest.sh -o install-latest.sh
```

```bash
chmod +x install-latest.sh
```

```bash
./install-latest.sh
```

--------------------------------

### Display CLI Help and Version

Source: https://docs.nvidia.com/brev/cli/cli-overview.md

Use these commands to verify the installation and explore available CLI functionality.

```bash
# Get help on any command
brev --help
brev <command> --help

# Check CLI version
brev --version
```

--------------------------------

### Install Brev CLI on macOS

Source: https://docs.nvidia.com/brev/cli/getting-started.md

Use Homebrew to install the CLI and verify the installation.

```bash
# Add the Brev tap and install
brew install brevdev/homebrew-brev/brev

# Verify installation
brev --version
```

--------------------------------

### Inspect Startup Logs

Source: https://docs.nvidia.com/brev/troubleshooting/common-issues.md

Displays the full initialization sequence to diagnose setup failures like repository cloning issues.

```bash
brev shell my-instance --host
sudo cat /var/log/brev-workspace.log
```

--------------------------------

### Install Brev CLI

Source: https://docs.nvidia.com/brev/getting-started/quickstart.md

Commands to install the Brev CLI on different operating systems and verify the installation.

```bash
# Install with Homebrew
brew install brevdev/homebrew-brev/brev

# Verify installation
brev --version
```

```bash
# Download and run the installer
bash -c "$(curl -fsSL https://raw.githubusercontent.com/brevdev/brev-cli/main/bin/install-latest.sh)"

# Verify installation
brev --version
```

```powershell
# First, install WSL with Ubuntu (run in PowerShell as admin)
wsl --install -d Ubuntu-22.04

# Restart your computer, then open Ubuntu and run:
bash -c "$(curl -fsSL https://raw.githubusercontent.com/brevdev/brev-cli/main/bin/install-latest.sh)"

# Verify installation
brev --version
```

--------------------------------

### Install Brev CLI

Source: https://docs.nvidia.com/brev/guides/registered-compute/register-dgx-spark.md

Install the Brev CLI on your DGX Spark system.

```bash
curl -fsSL https://brev.nvidia.com/install | bash
```

--------------------------------

### Common Instance Creation Examples

Source: https://docs.nvidia.com/brev/cli/instance-creation.md

Standard usage patterns for creating instances with specific hardware requirements, startup scripts, and cluster configurations.

```bash
# Simple creation with smart defaults
brev create my-instance

# Specific type with startup script
brev create my-instance --type g5.xlarge -s @setup.sh

# High-VRAM GPU with fallback
brev create training-box --gpu-name A100 --min-vram 40

# Create 3-node cluster
brev create my-cluster --count 3 --type g5.xlarge

# Preview what would be created
brev create my-instance --dry-run

# Full pipeline: search → create → setup → open
brev search --gpu-name A100 | brev create my-box | brev exec @setup.sh
```

--------------------------------

### Install the Brev agent skill via CLI

Source: https://docs.nvidia.com/brev/guides/ai-agents/cli-with-agents.md

Installs the agent skill using the Brev CLI.

```bash
brev agent-skill install
```

--------------------------------

### Install Required Dependencies

Source: https://docs.nvidia.com/brev/troubleshooting/instances-gpus/gpu-detection-pytorch.md

Install optional dependencies like xformers and torchao required by Unsloth.

```bash
pip install xformers torchao
```

--------------------------------

### View installed skill file structure

Source: https://docs.nvidia.com/brev/guides/ai-agents/cli-with-agents.md

Displays the directory structure of the installed Brev CLI skill.

```text
~/.claude/skills/brev-cli/
├── SKILL.md              # Skill definition and quick reference
├── prompts/
│   ├── quick-session.md  # Quick GPU session workflow
│   ├── ml-training.md    # ML training setup workflow
│   └── cleanup.md        # Instance cleanup workflow
├── reference/
│   ├── commands.md       # Full command reference
│   └── search-filters.md # GPU/CPU search filter details
└── examples/
    └── common-patterns.md # Common usage patterns
```

--------------------------------

### One-line installation command

Source: https://docs.nvidia.com/brev/troubleshooting/cli-installation/cli-installation-linux.md

The standard installation command that may fail silently in certain shell configurations.

```bash
bash -c "$(curl -fsSL https://raw.githubusercontent.com/brevdev/brev-cli/main/bin/install-latest.sh)"
```

--------------------------------

### Install NGC CLI

Source: https://docs.nvidia.com/brev/guides/inference-deployment/ngc-developer-guide.md

Commands to install the NGC CLI via pip or by downloading the binary directly.

```bash
# Using pip (recommended)
pip install ngc-cli

# Or download directly
wget -O ngc https://api.ngc.nvidia.com/v2/resources/nvidia/ngc-apps/ngc_cli/versions/3.31.0/files/ngccli_linux.zip
unzip ngccli_linux.zip
chmod +x ngc-cli/ngc
export PATH=$PATH:$(pwd)/ngc-cli
```

--------------------------------

### Verify CLI installation

Source: https://docs.nvidia.com/brev/troubleshooting/cli-installation/cli-installation-linux.md

Commands to confirm the binary exists, is in the system PATH, and is functional.

```bash
# Verify the binary exists
ls ~/.local/bin/brev

# Verify CLI is in PATH
which brev

# Verify CLI works
brev --help
```

--------------------------------

### Initialize NGC CLI and Pull Container

Source: https://docs.nvidia.com/brev/guides/inference-deployment/ngc-developer-guide.md

Install the CLI, configure authentication, and pull a PyTorch container image from the NGC registry.

```bash
# Get started in 3 commands:
pip install ngc-cli
ngc config set  # Enter your API key
docker pull nvcr.io/nvidia/pytorch:24.01-py3
```

--------------------------------

### Verify GPU Setup

Source: https://docs.nvidia.com/brev/guides/inference-deployment/deploying-nims

Runs a test container to confirm that the NVIDIA runtime and GPU drivers are correctly configured.

```bash
docker run --rm --runtime=nvidia --gpus all ubuntu nvidia-smi
```

--------------------------------

### Install the Brev agent skill via curl

Source: https://docs.nvidia.com/brev/guides/ai-agents/cli-with-agents.md

Installs the agent skill using a shell script via curl.

```bash
curl -fsSL https://raw.githubusercontent.com/brevdev/brev-cli/main/scripts/install-agent-skill.sh | bash
```

--------------------------------

### GPU search examples

Source: https://docs.nvidia.com/brev/cli/search-discovery.md

Common usage patterns for filtering and sorting GPU instance search results.

```bash
# List all available GPUs (sorted by price)
brev search

# Wide mode with extra columns
brev search gpu --wide

# Filter by GPU model
brev search --gpu-name A100
brev search --gpu-name H100

# Filter by VRAM
brev search --min-vram 40 --sort price

# Fast-booting GPUs
brev search --max-boot-time 3 --sort price

# Stoppable instances with minimum VRAM
brev search --stoppable --min-total-vram 40 --sort price
```

--------------------------------

### Revert to CPU-only Mode

Source: https://docs.nvidia.com/brev/troubleshooting/instances-gpus/gpu-detection-pytorch.md

Use this workaround if GPU/CUDA installation fails.

```bash
pip uninstall unsloth unsloth-zoo
pip install unsloth-cpu
```

--------------------------------

### CPU Search Examples

Source: https://docs.nvidia.com/brev/cli/search-discovery.md

Common usage patterns for filtering and sorting CPU instances, including piping results to the creation command.

```bash
# All CPU instances
brev search cpu

# AWS CPU instances
brev search cpu --provider aws

# High-memory instances
brev search cpu --min-ram 64 --sort price

# ARM architecture
brev search cpu --arch arm64

# Many-core instances
brev search cpu --min-vcpu 16 --sort price

# Pipe to create
brev search cpu --min-ram 64 | brev create my-cpu-box
```

--------------------------------

### Check installation exit code

Source: https://docs.nvidia.com/brev/troubleshooting/cli-installation/cli-installation-linux.md

Captures the exit status of the last command to identify if the installation failed.

```bash
echo $?
```

--------------------------------

### Verify API Key in VM Setup Script

Source: https://docs.nvidia.com/brev/concepts/launchables.md

Use this bash script to validate that a required environment variable is present during the VM initialization phase.

```bash
#!/usr/bin/env bash
set -euo pipefail

python3 - <<'PY'
import os

if not os.environ.get("API_KEY"):
    raise RuntimeError("API_KEY is required")

print("API_KEY is available during setup.")
# Perform one-time initialization that uses the API key here.
PY
```

--------------------------------

### Install CUDA-enabled PyTorch and Framework

Source: https://docs.nvidia.com/brev/troubleshooting/instances-gpus/gpu-detection-pytorch.md

Reinstall PyTorch with specific CUDA wheels if the current environment lacks GPU support.

```bash
pip uninstall -y torch torchvision torchaudio
pip install --index-url https://download.pytorch.org/whl/cu124 torch torchvision torchaudio
```

```bash
pip install -U unsloth unsloth-zoo
```

--------------------------------

### Natural language agent commands

Source: https://docs.nvidia.com/brev/guides/ai-agents/cli-with-agents.md

Examples of natural language prompts for managing instances via an AI agent.

```text
"Create an A100 instance for ML training"
"Search for GPUs with 40GB VRAM under $3/hr"
"Stop all my running instances"
"Open my-instance in Cursor"
"What GPUs are available with fast boot times?"
```

--------------------------------

### Manage Helm Charts

Source: https://docs.nvidia.com/brev/guides/inference-deployment/ngc-developer-guide.md

Commands to add the NGC Helm repository and install charts like the GPU Operator.

```bash
# Add NGC Helm repo
helm repo add nvidia https://helm.ngc.nvidia.com/nvidia

# Search for charts
helm search repo nvidia

# Install a chart (e.g., GPU Operator)
helm install gpu-operator nvidia/gpu-operator \
  --set driver.enabled=true \
  --namespace gpu-operator \
  --create-namespace
```

--------------------------------

### Customize Environment Packages

Source: https://docs.nvidia.com/brev/concepts/environments.md

Commands for installing Python and system-level packages, and pulling Docker images within the workspace.

```bash
# Navigate to workspace
cd /home/ubuntu/workspace

# Install Python packages
pip install transformers accelerate

# Install system packages
sudo apt update && sudo apt install -y htop

# Pull Docker images
docker pull nvcr.io/nvidia/pytorch:24.01-py3
```

--------------------------------

### Manage Docker Compose stacks

Source: https://docs.nvidia.com/brev/guides/development-tools/custom-containers.md

Commands to start and monitor multi-container applications defined in a docker-compose.yml file.

```bash
# Start the stack
docker compose up -d

# View logs
docker compose logs -f
```

--------------------------------

### Test skill changes from a branch

Source: https://docs.nvidia.com/brev/guides/ai-agents/cli-with-agents.md

Installs the skill from a specific branch for testing purposes.

```bash
BREV_SKILL_BRANCH=my-branch brev agent-skill install
```

```bash
curl -fsSL https://raw.githubusercontent.com/brevdev/brev-cli/main/scripts/install-agent-skill.sh | bash -s -- --branch my-branch
```

--------------------------------

### Start JupyterLab on Remote Instance

Source: https://docs.nvidia.com/brev/guides/development-tools/jupyter-notebooks.md

Connect to the instance and launch the JupyterLab server without opening a browser on the remote machine.

```bash
# Connect to your instance
brev shell my-instance

# Start JupyterLab (no browser on remote)
jupyter lab --no-browser --port=8888
```

--------------------------------

### Manage NIMs with NGC CLI

Source: https://docs.nvidia.com/brev/guides/inference-deployment/deploying-nims

Installs the NGC CLI, configures authentication, and lists available NIM images.

```bash
# Install NGC CLI
pip install ngc-cli

# Configure with your API key
ngc config set

# List available NIMs
ngc registry image list --format_type csv nvcr.io/nim/meta/*
```

--------------------------------

### Reinstall Torch in Correct Environment

Source: https://docs.nvidia.com/brev/troubleshooting/instances-gpus/gpu-detection-pytorch.md

Use this if PyTorch is installed in a different environment or missing from the PATH.

```bash
which python3
python3 -m pip install --force-reinstall torch
```

--------------------------------

### Register a machine with Brev

Source: https://docs.nvidia.com/brev/cli/registered-compute.md

Registers a Linux machine with the Brev organization. Requires sudo access and installs the NetBird VPN agent.

```bash
brev register -n <name>
```

```bash
# Register with a name
brev register -n "My DGX Spark"

# Register to a specific organization
brev register --name "Lab Workstation" --org my-team
```

--------------------------------

### Pull Containers from NGC

Source: https://docs.nvidia.com/brev/guides/inference-deployment/ngc-developer-guide.md

Examples of pulling various public and partner container images using the Docker CLI.

```bash
# Format: nvcr.io/<organization>/<repository>:<tag>

# NVIDIA's public containers
docker pull nvcr.io/nvidia/pytorch:24.01-py3
docker pull nvcr.io/nvidia/tensorflow:24.01-tf2-py3
docker pull nvcr.io/nvidia/tritonserver:24.01-py3

# Partner containers (ISV)
docker pull nvcr.io/isv-ngc-partner/company/container:version
```

--------------------------------

### brev register

Source: https://docs.nvidia.com/brev/cli/registered-compute.md

Registers a Linux machine with your Brev organization, installs the NetBird mesh VPN, and profiles the hardware.

```APIDOC
## brev register

### Description
Registers a Linux machine with your Brev organization. This installs the NetBird mesh VPN, profiles the machine's hardware, and registers it with the Brev API.

### Usage
`brev register -n <name>`

### Flags
- **--name** (`-n`) - Required - Display name for the machine.
- **--org** (`-o`) - Optional - Organization to register with (defaults to active org).
```

--------------------------------

### Quick Dev Session Initialization

Source: https://docs.nvidia.com/brev/cli/composable-workflows.md

Create a development environment and open it in Cursor in one command.

```bash
# One-liner: create and open in Cursor
brev create dev-box | brev open cursor
```

--------------------------------

### Create instance with smart defaults

Source: https://docs.nvidia.com/brev/cli/instance-creation.md

Uses default criteria to select the cheapest available GPU instance.

```bash
# Uses smart defaults — picks the cheapest GPU matching the above criteria
brev create my-instance
```

--------------------------------

### Configure startup scripts

Source: https://docs.nvidia.com/brev/cli/instance-creation.md

Executes commands or scripts upon instance boot using inline strings or file references.

```bash
brev create my-instance -s 'pip install torch'
```

```bash
brev create my-instance -s @setup.sh
brev create my-instance -s @./scripts/install-deps.sh
```

--------------------------------

### Search for development instances

Source: https://docs.nvidia.com/brev/cli/search-discovery.md

Filters for instances with a maximum boot time of 3 minutes, sorted by price.

```bash
brev search --max-boot-time 3 --sort price
```

--------------------------------

### Create and Connect to Instance

Source: https://docs.nvidia.com/brev/getting-started/quickstart.md

Provision a new GPU instance and open a remote shell session.

```bash
# Create a GPU instance with an NVIDIA L40S
brev create my-instance --gpu "nebius.l40sx1.pcie"

# Open a shell on your instance
brev shell my-instance
```

--------------------------------

### Initialize Workspace and Virtual Environment

Source: https://docs.nvidia.com/brev/concepts/environments.md

Standard commands to navigate to the persistent workspace and set up a Python virtual environment.

```bash
# Navigate to workspace
cd /home/ubuntu/workspace

# Clone your project
git clone https://github.com/your/repo.git

# Create a virtual environment
python3 -m venv .venv
source .venv/bin/activate
```

--------------------------------

### Create an instance with Brev CLI

Source: https://docs.nvidia.com/brev/cli/instance-creation.md

Basic syntax for initiating instance creation using the brev create command.

```bash
brev create [name] [flags]
```

--------------------------------

### Create empty instance

Source: https://docs.nvidia.com/brev/cli/instance-management.md

Initialize an instance without a linked Git repository.

```bash
brev start --name my-sandbox
```

--------------------------------

### Remove installed CLI

Source: https://docs.nvidia.com/brev/troubleshooting/cli-installation/cli-installation-linux.md

Deletes the Brev CLI binary from the local directory.

```bash
rm -f ~/.local/bin/brev
```

--------------------------------

### Provision a CPU instance

Source: https://docs.nvidia.com/brev/guides/ai-agents/agent-sandboxes.md

Create a new CPU instance for agent deployment using the Brev CLI.

```bash
brev create my-agent
```

--------------------------------

### Downloading training results

Source: https://docs.nvidia.com/brev/guides/development-tools/file-transfer-scp.md

Retrieves output directories from the instance.

```bash
brev copy -r my-instance:/home/ubuntu/workspace/outputs ./local-outputs
```

--------------------------------

### Create instance from Git URL

Source: https://docs.nvidia.com/brev/cli/instance-management.md

Initialize an instance from a remote repository with optional naming and organization scoping.

```bash
# Create from GitHub
brev start https://github.com/user/repo.git

# With custom name
brev start https://github.com/user/repo.git --name my-project

# In a specific organization
brev start https://github.com/user/repo.git --org my-team
```

--------------------------------

### Preview Instance Creation with Dry Run

Source: https://docs.nvidia.com/brev/cli/instance-creation.md

Use the --dry-run flag to validate instance configurations without provisioning resources.

```bash
brev create my-instance --dry-run
brev create my-instance --gpu-name A100 --dry-run
```

--------------------------------

### Reinstall PyTorch with CUDA Support

Source: https://docs.nvidia.com/brev/troubleshooting/instances-gpus/gpu-detection-pytorch.md

Fixes issues where PyTorch is installed without CUDA support.

```bash
pip uninstall -y torch torchvision torchaudio
pip install --index-url https://download.pytorch.org/whl/cu124 \
    torch torchvision torchaudio
```

--------------------------------

### Create instance with GPU

Source: https://docs.nvidia.com/brev/cli/instance-management.md

Specify the GPU type and count during instance creation.

```bash
# Default GPU (T4)
brev start https://github.com/user/repo.git

# Specify GPU type
brev start https://github.com/user/repo.git --gpu "n1-highmem-4:nvidia-tesla-a100:1"
```

--------------------------------

### Search for production instances

Source: https://docs.nvidia.com/brev/cli/search-discovery.md

Filters for A100 GPU instances with a maximum boot time of 5 minutes, sorted by price.

```bash
brev search --gpu-name A100 --max-boot-time 5 --sort price
```

--------------------------------

### Configure NGC Authentication

Source: https://docs.nvidia.com/brev/guides/inference-deployment/ngc-developer-guide.md

Methods for authenticating the NGC CLI using interactive setup or environment variables.

```bash
# Interactive setup
ngc config set

# Or set environment variables
export NGC_API_KEY=<your-api-key>
export NGC_ORG=<your-org>  # Optional: for enterprise users
```

--------------------------------

### Syncing project files

Source: https://docs.nvidia.com/brev/guides/development-tools/file-transfer-scp.md

Uploads an entire project directory to the instance workspace.

```bash
brev copy -r ./my-ml-project my-instance:/home/ubuntu/workspace/
```

--------------------------------

### Search for instances

Source: https://docs.nvidia.com/brev/cli/search-discovery.md

Basic command structure for searching GPU or CPU instances.

```bash
brev search [gpu|cpu] [flags]
```

--------------------------------

### List instances and organizations

Source: https://docs.nvidia.com/brev/cli/instance-management.md

View workspaces and organizations using the list command.

```bash
# List instances in current organization
brev list

# List all organizations you belong to
brev list orgs

# List instances in a specific organization
brev list --org my-team
```

--------------------------------

### Access Instance Shell

Source: https://docs.nvidia.com/brev/troubleshooting/instances-gpus/instance-stuck-in-waiting-state.md

Connect to a running instance via the CLI.

```bash
brev shell <instance-name>
```

--------------------------------

### Search for budget-conscious instances

Source: https://docs.nvidia.com/brev/cli/search-discovery.md

Filters for instances with at least 16GB of VRAM, sorted by price, and limits output to the top 5 results.

```bash
brev search --min-vram 16 --sort price | head -5
```

--------------------------------

### Connect to the instance

Source: https://docs.nvidia.com/brev/guides/ai-agents/agent-sandboxes.md

Establish a shell connection to the running agent instance.

```bash
brev shell my-agent
```

--------------------------------

### Reset Instance State

Source: https://docs.nvidia.com/brev/cli/advanced-commands.md

Clears the container environment while preserving persistent storage. Useful for resolving environment corruption or starting fresh.

```bash
brev reset my-instance
```

--------------------------------

### Connect to Machine

Source: https://docs.nvidia.com/brev/guides/registered-compute/managing-registered-compute.md

Use this command from a local machine to connect to a registered instance after access has been granted.

```bash
brev shell <machine-name>
```

--------------------------------

### Create instance clusters

Source: https://docs.nvidia.com/brev/cli/instance-creation.md

Deploys multiple instances simultaneously using the count flag.

```bash
brev create my-cluster --count 3 --type g5.xlarge
```

--------------------------------

### Chaining create and access commands

Source: https://docs.nvidia.com/brev/cli/composable-workflows.md

Combine search, creation, and execution commands into a single pipeline.

```bash
# Search, create, and open
brev search --gpu-name A100 | brev create my-box | brev open cursor

# Create and run a command
brev create my-box | brev exec "nvidia-smi"

# Create and run a setup script
brev create my-box | brev exec @setup.sh

# Create cluster and verify all nodes
brev create my-cluster --count 3 | brev exec "hostname && nvidia-smi"
```

--------------------------------

### Test GPU Detection via Python

Source: https://docs.nvidia.com/brev/troubleshooting/instances-gpus/gpu-detection-pytorch.md

Run a quick command to verify Unsloth detects the GPU.

```bash
python -c "import unsloth; print('GPU available:', unsloth.torch.cuda.is_available())"
```

--------------------------------

### Open VS Code on instance

Source: https://docs.nvidia.com/brev/guides/development-tools/vscode-setup.md

Use the Brev CLI to automatically configure and open the VS Code environment for a specific instance.

```bash
brev open my-instance
```

--------------------------------

### brev create

Source: https://docs.nvidia.com/brev/cli/instance-creation.md

Creates a new GPU or CPU instance using the Brev CLI with support for various configuration flags and search filters.

```APIDOC
## brev create

### Description
Creates GPU and CPU instances with smart defaults, fallback chains, startup scripts, and composable piping.

### Usage
`brev create [name] [flags]`

### Aliases
`provision`, `gpu-create`, `gpu-retry`, `gcreate`

### Flags
- **--name** (-n) - Instance name.
- **--type** (-t) - Comma-separated instance types to try (fallback chain).
- **--count** (-c) - Number of instances to create (default: 1).
- **--parallel** (-p) - Parallel creation attempts (default: 1).
- **--startup-script** (-s) - Script to run on boot (inline string or @filepath).
- **--timeout** - Seconds to wait for ready (default: 300).
- **--detached** (-d) - Don't wait for instance to be ready.
- **--dry-run** - Show matching instance types without creating.

### Search Filter Flags
- **--gpu-name** (-g) - Filter by GPU name (partial match).
- **--provider** - Filter by cloud provider.
- **--min-vram** (-v) - Minimum VRAM per GPU (GB).
- **--min-total-vram** - Minimum total VRAM (GB).
- **--min-capability** - Minimum GPU compute capability.
- **--min-disk** - Minimum disk size (GB).
- **--max-boot-time** - Maximum boot time (minutes).
- **--stoppable** - Only stoppable instances.
- **--rebootable** - Only rebootable instances.
- **--flex-ports** - Only instances with configurable firewall.
- **--sort** - Sort by: price, vram, boot-time, etc.
- **--desc** - Sort descending.
```

--------------------------------

### Search for latest architecture instances

Source: https://docs.nvidia.com/brev/cli/search-discovery.md

Filters for instances with a minimum compute capability of 9.0, sorted by price.

```bash
brev search --min-capability 9.0 --sort price
```

--------------------------------

### Pipe search results to instance creation

Source: https://docs.nvidia.com/brev/cli/search-discovery.md

Piping output to another command allows for direct chaining, such as creating an instance based on a specific GPU name.

```bash
brev search --gpu-name A100 | brev create my-instance
```

--------------------------------

### Open instance in an editor

Source: https://docs.nvidia.com/brev/cli/connectivity.md

Opens the instance in a specified editor or sets a default editor for future sessions.

```bash
# Open in VS Code (default)
brev open my-instance

# Open in specific editor
brev open my-instance cursor
brev open my-instance windsurf
brev open my-instance tmux

# Set default editor
brev open --set-default cursor
```

--------------------------------

### Connect via SSH

Source: https://docs.nvidia.com/brev/cli/connectivity.md

Connect to your instance using the Brev-managed key or via SSH configuration. Manual port forwarding can also be performed using standard SSH flags.

```bash
# Connect using instance name (configured by brev refresh)
ssh my-instance

# Or use the full SSH command
ssh -i ~/.brev/brev.pem ubuntu@<instance-ip>

# Manual port forward with SSH
ssh -i ~/.brev/brev.pem -p 22 -L LOCAL_PORT:localhost:REMOTE_PORT ubuntu@<instance-ip>
```

--------------------------------

### Search for ML training instances

Source: https://docs.nvidia.com/brev/cli/search-discovery.md

Filters for instances with at least 40GB of VRAM, sorted by price.

```bash
brev search --min-vram 40 --sort price
```

--------------------------------

### Filter instances during creation

Source: https://docs.nvidia.com/brev/cli/instance-creation.md

Uses search filters or piped output to select instances based on hardware requirements.

```bash
brev create my-instance --gpu-name A100 --min-vram 40
```

```bash
brev search --gpu-name A100 | brev create my-instance
```

--------------------------------

### Connect to Brev Instance

Source: https://docs.nvidia.com/brev/guides/inference-deployment/deploying-nims

Establishes a shell session with the specified Brev instance.

```bash
brev shell my-instance
```

--------------------------------

### Verify GPU Access

Source: https://docs.nvidia.com/brev/getting-started/quickstart.md

Check that the GPU is correctly configured and accessible within the instance.

```bash
nvidia-smi
```

--------------------------------

### Open Brev instance in VS Code

Source: https://docs.nvidia.com/brev/troubleshooting/ide-connectivity/vscode-windows-wsl.md

Command used to initiate a connection to a Brev instance from the terminal.

```bash
brev open <instance_name> code
```

--------------------------------

### Troubleshoot SSH Connection Issues

Source: https://docs.nvidia.com/brev/troubleshooting/common-issues.md

Commands to verify instance status and refresh SSH configuration when connection is refused.

```bash
# Check instance status
brev list

# Update SSH config
brev refresh

# Try connecting again
brev shell my-instance
```

--------------------------------

### Create Instance in Detached Mode

Source: https://docs.nvidia.com/brev/cli/instance-creation.md

Use the --detached flag to initiate creation without waiting for the instance to reach a ready state.

```bash
brev create my-instance --detached
```

--------------------------------

### Troubleshoot Login Failures

Source: https://docs.nvidia.com/brev/troubleshooting/common-issues.md

Alternative login methods for browser authentication issues or headless environments.

```bash
# Skip auto-opening browser
brev login --skip-browser

# Token-based login for headless/CI
brev login --token
```

--------------------------------

### Troubleshooting instance status

Source: https://docs.nvidia.com/brev/guides/development-tools/file-transfer-scp.md

Commands to verify instance status and refresh configuration.

```bash
brev refresh
brev ls  # Check instance status
```

```bash
brev ls
```

--------------------------------

### Search CPU Instances Command

Source: https://docs.nvidia.com/brev/cli/search-discovery.md

Basic command structure for searching CPU-only instances.

```bash
brev search cpu [flags]
```

--------------------------------

### Manage Resources

Source: https://docs.nvidia.com/brev/guides/inference-deployment/ngc-developer-guide.md

Commands to list and download datasets or other resources from the NGC registry.

```bash
# List resources
ngc registry resource list

# Download a resource
ngc registry resource download-version nvidia/tao/tao_getting_started:1.0
```

--------------------------------

### Run and Verify GPU in Docker

Source: https://docs.nvidia.com/brev/troubleshooting/common-issues.md

Use these commands to enable GPU access for a container and confirm visibility via nvidia-smi.

```bash
# Run with GPU access
docker run --gpus all -it your-image

# Verify GPU in container
nvidia-smi
```

--------------------------------

### Chain Commands with Output Piping

Source: https://docs.nvidia.com/brev/cli/instance-creation.md

Pipe the instance name output from brev create to subsequent commands for automated workflows.

```bash
# Create and open in an editor
brev create my-instance | brev open cursor

# Create and run a command
brev create my-instance | brev exec "nvidia-smi"

# Create and run a setup script
brev create my-instance | brev exec @setup.sh

# Full pipeline: search, create, open
brev search --gpu-name A100 | brev create my-box | brev open cursor
```

--------------------------------

### Run NIM Container

Source: https://docs.nvidia.com/brev/guides/inference-deployment/deploying-nims

Executes the NIM container with necessary GPU, memory, and volume configurations.

```bash
docker run -it --rm --name=$CONTAINER_NAME \
  --runtime=nvidia \
  --gpus all \
  --shm-size=16GB \
  -e NGC_CLI_API_KEY \
  -v "$LOCAL_NIM_CACHE:/opt/nim/.cache" \
  -u $(id -u) \
  -p 8000:8000 \
  $IMG_NAME
```

--------------------------------

### Connect to DGX Spark

Source: https://docs.nvidia.com/brev/guides/registered-compute/register-dgx-spark.md

Initiate a shell session to the registered DGX Spark.

```bash
brev shell my-dgx-spark
```

--------------------------------

### Authenticate with Token

Source: https://docs.nvidia.com/brev/troubleshooting/cli-installation/cli-ssh-authentication-macos.md

Use the authentication token from the Brev dashboard to log in and verify the connection.

```bash
brev login --token
```

```bash
brev ls
```

--------------------------------

### Search for GPU instances

Source: https://docs.nvidia.com/brev/cli/search-discovery.md

Command structure for searching GPU instances with optional flags.

```bash
brev search [flags]
brev search gpu [flags]
```

--------------------------------

### brev open

Source: https://docs.nvidia.com/brev/cli/connectivity.md

Opens the instance in a specified editor or environment such as VS Code, Cursor, Windsurf, or tmux.

```APIDOC
## brev open

### Description
Open your instance in VS Code, Cursor, Windsurf, or tmux.

### Usage
`brev open <instance-name> [editor]`

### Supported Editors
- **VS Code** - `brev open my-instance` or `brev open my-instance code`
- **Cursor** - `brev open my-instance cursor`
- **Windsurf** - `brev open my-instance windsurf`
- **tmux** - `brev open my-instance tmux`

### Flags
- **--host** - SSH into the host machine instead of the container.
- **--wait / -w** - Wait for setup to finish before opening.
- **--dir / -d** - Specify directory to open.
- **--set-default** - Set default editor (code, cursor, windsurf, tmux)
```

--------------------------------

### Specify instance types

Source: https://docs.nvidia.com/brev/cli/instance-creation.md

Defines specific instance types or a fallback chain for creation.

```bash
brev create my-instance --type g5.xlarge
```

```bash
brev create my-instance --type g5.xlarge,g5.2xlarge,g5.4xlarge
```

--------------------------------

### Embed a Launchable badge in Markdown

Source: https://docs.nvidia.com/brev/concepts/launchables.md

Use this markdown snippet to add a clickable deployment button to your README or documentation files.

```markdown
[![Launch on Brev](https://brev-assets.s3.us-west-1.amazonaws.com/nv-lb-dark.svg)](https://brev.nvidia.com/launchable/deploy?launchableID=your-launchable-id)
```

--------------------------------

### Run GPU-enabled Docker containers via CLI

Source: https://docs.nvidia.com/brev/guides/development-tools/custom-containers.md

Use the --gpus all flag to enable GPU access for Docker containers on Brev instances.

```bash
# Run PyTorch container with GPU
docker run --gpus all -it nvcr.io/nvidia/pytorch:24.01-py3 python -c "import torch; print(torch.cuda.is_available())"

# Run with port mapping
docker run --gpus all -p 8888:8888 -it nvcr.io/nvidia/pytorch:24.01-py3
```

--------------------------------

### List instances in interactive mode

Source: https://docs.nvidia.com/brev/cli/composable-workflows.md

Displays a formatted table with headers and colors when output is directed to a terminal.

```bash
brev ls
```

--------------------------------

### Download files using brev copy

Source: https://docs.nvidia.com/brev/guides/development-tools/file-transfer-scp.md

Downloads files or directories from a Brev instance to the local machine.

```bash
brev copy my-instance:/home/ubuntu/results.csv ./
```

```bash
brev copy -r my-instance:/home/ubuntu/outputs ./local-outputs
```

--------------------------------

### Manage Organization Context

Source: https://docs.nvidia.com/brev/cli/advanced-commands.md

Use these commands to list available organizations or switch the active organization context for subsequent operations.

```bash
# List available organizations
brev set

# Switch to a specific organization
brev set my-org
```

--------------------------------

### Download Models via NGC CLI

Source: https://docs.nvidia.com/brev/guides/inference-deployment/ngc-developer-guide.md

Commands to list and download models from the NGC registry to a local directory.

```bash
# List available models
ngc registry model list

# Download a specific model
ngc registry model download-version nvidia/nemo/megatron_gpt_345m:1.0

# Download to specific directory
ngc registry model download-version nvidia/nemo/megatron_gpt_345m:1.0 --dest ./models/
```

--------------------------------

### Preview and Execute Destructive Operations

Source: https://docs.nvidia.com/brev/cli/composable-workflows.md

Verify instance lists before applying destructive commands like stop or delete.

```bash
# Preview what will be stopped
brev ls | awk '/RUNNING/ {print $1}'

# Then actually stop
brev ls | awk '/RUNNING/ {print $1}' | brev stop
```

--------------------------------

### Authenticate with Brev

Source: https://docs.nvidia.com/brev/cli/advanced-commands.md

Opens a browser window for authentication or refreshes existing credentials.

```bash
brev login
```

--------------------------------

### Expose NIM via Port Forwarding

Source: https://docs.nvidia.com/brev/guides/inference-deployment/deploying-nims

Use this command to map the local machine port to the instance port for direct API access.

```bash
# From your local machine
brev port-forward my-instance --port 8000:8000

# Access at localhost:8000
curl -X POST http://localhost:8000/v1/chat/completions ...
```

--------------------------------

### Chain instance listing and stopping

Source: https://docs.nvidia.com/brev/cli/composable-workflows.md

Pipes the output of instance listing directly into the stop command for batch processing.

```bash
brev ls | brev stop
```

--------------------------------

### Search for large model training instances

Source: https://docs.nvidia.com/brev/cli/search-discovery.md

Filters for instances with at least 80GB of total VRAM, sorted by price.

```bash
brev search --min-total-vram 80 --sort price
```

--------------------------------

### Display search results as a table

Source: https://docs.nvidia.com/brev/cli/search-discovery.md

The default output format for interactive sessions displays instance specifications in a table.

```text
TYPE            GPU         COUNT  VRAM   TOTAL   $/HR   BOOT  FEATURES
g5.xlarge       A10G        1      24GB   24GB    $1.00  2m    S R P
g5.2xlarge      A10G        1      24GB   24GB    $1.20  2m    S R P
p4d.24xlarge    A100        8      40GB   320GB   $32.77 5m    S R
```

--------------------------------

### Retrieve instance data as JSON

Source: https://docs.nvidia.com/brev/cli/composable-workflows.md

Outputs instance information in a structured JSON format for programmatic parsing.

```bash
brev ls --json
```

--------------------------------

### Performing batch operations

Source: https://docs.nvidia.com/brev/cli/composable-workflows.md

Pass multiple instance names directly to commands for bulk actions.

```bash
# Stop multiple instances
brev stop instance1 instance2 instance3

# Delete multiple instances
brev delete instance1 instance2 instance3

# Start multiple instances
brev start instance1 instance2

# Run on multiple instances
brev exec instance1 instance2 instance3 "nvidia-smi"
```

--------------------------------

### Stopping all instances

Source: https://docs.nvidia.com/brev/cli/composable-workflows.md

Use the --all flag to target every running instance.

```bash
brev stop --all
```

--------------------------------

### Upload files using brev copy

Source: https://docs.nvidia.com/brev/guides/development-tools/file-transfer-scp.md

Uploads local files or directories to a Brev instance.

```bash
brev copy ./local-file.txt my-instance:/home/ubuntu/
```

```bash
brev copy -r ./my-project my-instance:/home/ubuntu/
```

--------------------------------

### Filter GPU instances by compute capability

Source: https://docs.nvidia.com/brev/cli/search-discovery.md

Use the --min-capability flag to filter available instances based on their architecture generation.

```bash
# Ampere or newer
brev search --min-capability 8.0

# Hopper only
brev search --min-capability 9.0
```

--------------------------------

### Configure Default Editor

Source: https://docs.nvidia.com/brev/troubleshooting/common-issues.md

Set the default editor for the brev open command.

```bash
# Set a different default editor
brev open --set-default cursor
```

--------------------------------

### Multi-Node Cluster Orchestration

Source: https://docs.nvidia.com/brev/cli/composable-workflows.md

Provision multiple nodes and execute commands across the entire cluster.

```bash
# Create 3 nodes, install deps on all, verify
brev create my-cluster --count 3 --type g5.xlarge \
  | brev exec @install-deps.sh \
  | brev exec "nvidia-smi"
```

--------------------------------

### Manage Private Registry and Access

Source: https://docs.nvidia.com/brev/guides/inference-deployment/ngc-developer-guide.md

Commands to tag and push containers to a private registry and assign user roles for access control.

```bash
# Push to private registry
docker tag my-container:latest nvcr.io/my-org/my-team/my-container:v1.0
docker push nvcr.io/my-org/my-team/my-container:v1.0

# Manage access
ngc org user add --org my-org --email colleague@company.com --role REGISTRY_USER
```

--------------------------------

### Verify GPU and CUDA Availability

Source: https://docs.nvidia.com/brev/concepts/environments.md

Use these commands to check driver status and verify PyTorch GPU integration.

```bash
# Quick check - GPU and CUDA availability
nvidia-smi
nvcc --version

# Framework verification (requires PyTorch installed)
pip install torch  # if not using a PyTorch base image
python3 -c "import torch; print(f'CUDA available: {torch.cuda.is_available()}')"
python3 -c "import torch; print(f'GPU: {torch.cuda.get_device_name(0)}')"
```

--------------------------------

### Authenticate with Brev

Source: https://docs.nvidia.com/brev/cli/getting-started.md

Manage user authentication sessions using standard, token-based, or headless login methods.

```bash
# Standard login (opens browser)
brev login

# Login with token (for headless/CI environments)
brev login --token

# Skip browser auto-open
brev login --skip-browser
```

--------------------------------

### Connect to registered compute

Source: https://docs.nvidia.com/brev/cli/registered-compute.md

Uses shell or port-forwarding to connect to registered compute nodes.

```bash
# SSH into a registered machine
brev shell my-dgx-spark

# Forward a port from a registered machine
brev port-forward my-dgx-spark --port 8888:8888
```

--------------------------------

### Pull and Run Containers

Source: https://docs.nvidia.com/brev/guides/inference-deployment/ngc-developer-guide.md

Commands to pull a container image from the NGC registry and execute it with GPU support.

```bash
# Pull PyTorch container optimized for GPUs
docker pull nvcr.io/nvidia/pytorch:24.01-py3

# Run it
docker run --gpus all -it nvcr.io/nvidia/pytorch:24.01-py3
```