# CLI reference

Exact commands, arguments, and help text from the installed PLLM CLI.

[View canonical HTML](https://pllm.run/cli/reference/)

Document ID: `pllm.docs.reference.cli`  
Release: `0.1.0`

Use the CLI to run private inference, operate provider roles, benchmark the real transport, inspect configuration, and compare components. Open [Private inference](/cli/private-inference/), [Provider roles](/cli/provider-roles/), [Benchmarking](/cli/benchmarking/), or [Inspect components](/cli/inspect-and-research/). Choose a command group below for exact generated arguments.

## Complete grammar

```text
pllm gateway [OPTIONS]
pllm config show [OPTIONS]
pllm config export [OPTIONS]
pllm components list [OPTIONS]
pllm components show [OPTIONS]
pllm serve inference [OPTIONS]
pllm serve preparation [OPTIONS]
pllm benchmark run [OPTIONS]
pllm dev dashboard [OPTIONS]
```

Global options accepted at each parser level are `--format {human,json,jsonl}`, `--quiet`, `--no-color`, `--no-input`, and `--dry-run`. Root also accepts `--version`. Argument abbreviation is disabled.

[Download exact complete help](/downloads/cli-help.txt).

## `pllm --help`

```text
usage: pllm [-h] [--format {human,json,jsonl}] [--quiet] [--no-color]
            [--no-input] [--dry-run] [--version]
            COMMAND ...

Inspect PLLM configuration, run benchmarks, and operate private inference

positional arguments:
  COMMAND
    config              inspect or export public configuration
    components          inspect built-in component descriptors
    gateway             run the trusted local Responses API gateway
    serve               run an inference or preparation role
    benchmark           run reproducible local benchmarks
    dev                 development tools

options:
  -h, --help            show this help message and exit
  --format {human,json,jsonl}
                        result format (default: human)
  --quiet               suppress non-error diagnostics
  --no-color            disable colored output
  --no-input            fail instead of prompting
  --dry-run             report work without persistent writes or service
                        startup
  --version             show program's version number and exit
```