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

# NVIDIA Fleet Intelligence Client

`nvfleetint` provides command-line and Go interfaces to NVIDIA Fleet
Intelligence.

- **CLI:** Use the `nvfleetint` executable to inspect fleet status, node health,
  inventory, alerts, events, and reports from a terminal or automation
  workflow.
- **Go SDK:** Import
  `github.com/NVIDIA/fleet-intelligence-client/nvfleetint` to integrate Fleet
  Intelligence data into Go applications and services.

Both interfaces support fleet summaries, compute zones, node groups, nodes,
health history, alerts, events, tags, inventory reports, and error reports.

## Install

### CLI

#### Linux and macOS

Install `nvfleetint`:

```bash
curl -fsSL https://github.com/NVIDIA/fleet-intelligence-client/releases/latest/download/install.sh | bash
```

The installer verifies the release checksum and, on macOS, the Developer ID
signature and Apple notarization ticket before installing the executable.
For publisher-authenticated verification using the signed checksum manifest,
see [Verify release artifacts](/fleet-intel/client/verify-releases).

#### Windows

Install `nvfleetint`:

```powershell
Invoke-WebRequest https://github.com/NVIDIA/fleet-intelligence-client/releases/latest/download/install.ps1 -OutFile install.ps1
.\install.ps1
```

### Go SDK

Add the `nvfleetint` package to a Go module:

```bash
go get github.com/NVIDIA/fleet-intelligence-client/nvfleetint
```

See the [Go SDK guide](/fleet-intel/client/sdk) for client setup and usage.

## Quick Start

Choose an NGC API key:

- [Personal API key](https://docs.nvidia.com/ngc/latest/ngc-user-guide.html#generating-a-personal-api-key)
  for individual use.
- [Service key](https://org.ngc.nvidia.com/identity-access/service-keys) for
  programmatic integrations.

```bash
nvfleetint auth add     # prompts for the API key and API URL
nvfleetint overview
nvfleetint node list
```

`auth add` reads the key from stdin rather than a flag, so it never reaches
your shell history. The key is not echoed as you type it, and the API URL is
offered with the production endpoint already filled in — press Enter to take
it.

Credentials are stored in named profiles; with no name, `auth add` uses one
called `default`, so single-tenant setup is the one command above. To work
against several tenants or endpoints, name them: add more with `nvfleetint auth
add <name>`, switch the default with `nvfleetint auth use <name>`, and select one
for a single command with `--profile`. Re-running `auth add` with an existing
name changes that profile in place, which is how a key is rotated — press Enter
at the key prompt to keep the stored key without ever displaying it (`--yes`
covers replacing it from a script, where nothing can be confirmed):

```bash
nvfleetint auth add dev
nvfleetint node list --profile dev
```

Run `nvfleetint <command> --help` for command-specific flags.

## AI Skills

Install the bundled Fleet Intelligence skills for Claude Code, Cursor, and
Codex:

```bash
npx skills add NVIDIA/fleet-intelligence-client \
  --skill '*' \
  --agent claude-code --agent cursor --agent codex
```

Install and authenticate `nvfleetint` first. See
[Agent Skills](/fleet-intel/client/skills) for available skills and usage guidance.