---
# Code generated by go run ./scripts/gendocs. DO NOT EDIT.
title: "CLI reference"
description: "Manage vLLM on one GPU machine or several"
---

<!-- DO NOT EDIT | GENERATED CONTENT -->

## Usage

```console
vllm-helper
```

## Description

```console
vllm-helper -- serve large models with vLLM on one GPU machine or
several (models, switching, sharding, replica pools, recipes, optional
Coder Agents sync). Per machine, only one native model is loaded at a time.
```

## Subcommands

| Name                                  | Purpose                                                                                |
| ------------------------------------- | -------------------------------------------------------------------------------------- |
| [`adopt`](cli/adopt.md)               | Detect an already-running vLLM server and save it as a managed profile                 |
| [`coder-sync`](cli/coder-sync.md)     | Manually sync a profile's Coder model entry                                            |
| [`doctor`](cli/doctor.md)             | Diagnose hardware, running servers, integrations, and config drift                     |
| [`edit`](cli/edit.md)                 | Open a profile in $EDITOR and validate the result                                      |
| [`inspect`](cli/inspect.md)           | Show a model's license, weight size against this machine, and detected flags           |
| [`install-gguf`](cli/install-gguf.md) | Install the plugin that lets vLLM load GGUF models                                     |
| [`list`](cli/list.md)                 | List saved model profiles                                                              |
| [`logs`](cli/logs.md)                 | Show the server log, following by default on a terminal                                |
| [`new`](cli/new.md)                   | Pull a new HuggingFace model, save it as a profile, and switch to it                   |
| [`node`](cli/node.md)                 | Register machines by HTTP URL and SSH target                                           |
| [`pool`](cli/pool.md)                 | Run a router in front of replicas serving the same model                               |
| [`recipes`](cli/recipes.md)           | Tested launch configurations for known-good models                                     |
| [`remove`](cli/remove.md)             | Delete a profile, and its Coder model entry when sync is configured                    |
| [`restart`](cli/restart.md)           | Re-apply the active profile and restart the server                                     |
| [`search`](cli/search.md)             | Search HuggingFace and pick a model to pull and switch to                              |
| [`serve`](cli/serve.md)               | Run a web panel for this machine                                                       |
| [`set`](cli/set.md)                   | Change profile settings from the command line                                          |
| [`setup`](cli/setup.md)               | First-time setup: detect hardware and vLLM, write config, offer to install the service |
| [`shard`](cli/shard.md)               | Split one model across machines (Ray or vLLM multiprocessing)                          |
| [`spark`](cli/spark.md)               | DGX Spark helpers                                                                      |
| [`status`](cli/status.md)             | Server state, active model, and whether the endpoint responds                          |
| [`stop`](cli/stop.md)                 | Stop the server, keeping profiles and config                                           |
| [`switch`](cli/switch.md)             | Switch to a saved profile, restarting the server                                       |
| [`upgrade`](cli/upgrade.md)           | Update vllm-helper itself from a published release                                     |
| [`version`](cli/version.md)           | Print the version                                                                      |