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

# Client Setup

> Installation steps for the NVIDIA CCluster client

We provide a client library to programmatically interact with NVIDIA CCluster either through [command line](/ccluster/clients/setup) or using [Python SDK](/ccluster/clients/sdk).

Use the following command to install centml client.

```bash showLineNumbers={false}
pip install centml
```

Run the following command to log in using your NVIDIA CCluster client credentials.

```bash showLineNumbers={false}
centml login
```

The first time, you'll be prompted to log in to your CCluster account through a browser. Once logged in, you are all set to use the client. You can use the `centml` CLI command to list your deployments.

For **unattended automation** (CI pipelines, scheduled jobs, or integrations), use a [service account](/ccluster/resources/service-accounts) instead of browser login. Set `CENTML_SERVICE_ACCOUNT_ID` and `CENTML_SERVICE_ACCOUNT_SECRET` from the Client ID and Client Secret shown in the UI.

```bash showLineNumbers={false}
centml cluster ls

# Alternatively, as a shorthand
ccluster ls
```

Check out other supported commands:

```bash showLineNumbers={false}
ccluster --help
Usage: ccluster [OPTIONS] COMMAND [ARGS]...

  CentML cluster CLI tool

Options:
  --help  Show this message and exit.

Commands:
  capacity  Show GPU capacity across clusters
  delete    Delete a deployment
  exec      Execute a command in a deployment pod
  get       Get deployment details
  ls        List all deployments
  pause     Pause a deployment
  resume    Resume a deployment
  shell     Open an interactive shell to a deployment pod
```

| Command                       | Usage                                                                                                                 |
| ----------------------------- | --------------------------------------------------------------------------------------------------------------------- |
| `ls`                          | `ccluster ls [inference\|cserve\|compute\|rag\|job]` — list deployments, optionally filtered by type                  |
| `get`                         | `ccluster get <type> <id>` — show details for one deployment                                                          |
| `pause` / `resume` / `delete` | `ccluster pause <id>` — change or remove a deployment                                                                 |
| `shell`                       | `ccluster shell <id> [--pod NAME] [--shell bash\|sh\|zsh] [--first-pod]` — open an interactive shell in a running pod |
| `exec`                        | `ccluster exec <id> [--pod NAME] [--first-pod] -- <command>` — run a single command in a running pod                  |
| `capacity`                    | `ccluster capacity [--cluster-id ID]` — show GPU capacity across clusters                                             |

> **Note**
>
> `ccluster ls` does not list NVIDIA Dynamo deployments. View and manage Dynamo deployments from the console.

To logout, use the following command.

```bash showLineNumbers={false}
centml logout
```

## What's next

#### [Managing Vault Objects](/ccluster/resources/vault)

Learn how to generate and store NVIDIA CCluster tokens and other vault objects.

#### [Python SDK Reference](/ccluster/clients/sdk)

Learn how to access the NVIDIA CCluster using Python.