Getting Started

Install and configure the VLM Run CLI

The VLM Run CLI (vlmrun) lets you interact with the VLM Run platform directly from your terminal — chat with Orion, generate structured predictions, manage files and skills, and more.

Installation

The CLI is included with the VLM Run Python SDK. For the full experience with rich terminal output, install with the cli extra:

# Basic installation
pip install vlmrun

# Full CLI with rich terminal output (recommended)
pip install vlmrun[cli]

Verify the installation:

vlmrun --version

Configuration

Quick Setup

Initialize a config file and set your API key:

vlmrun config init
vlmrun config set --api-key "your-api-key"

Configuration is stored at ~/.vlmrun/config.toml.

Environment Variables

Alternatively, set your API key via environment variables:

export VLMRUN_API_KEY="your-api-key"
export VLMRUN_BASE_URL="https://api.vlm.run/v1"  # Optional

Managing Config

# Show current configuration
vlmrun config show

# Set values
vlmrun config set --api-key "your-api-key"
vlmrun config set --base-url "https://api.vlm.run/v1"

# Unset values
vlmrun config unset --api-key
vlmrun config unset --base-url

# Re-initialize (overwrite existing)
vlmrun config init --force

Global Options

All commands accept these global options:

Option	Description
--api-key TEXT	API key (overrides config/env)
--base-url TEXT	API base URL (overrides config/env)
--debug	Enable debug mode
-v, --version	Show version and exit
--help	Show help and exit

Commands

Command	Description
chat	Chat with Orion to process images, videos, and documents
execute	Execute agents with files, skills, and structured output
generate	Generate structured predictions from files
executions	List and retrieve agent execution results
predictions	List and retrieve prediction results
files	Upload, list, retrieve, and delete files
skills	Create, list, lookup, update, and download skills
hub	Browse available domains and JSON schemas
models	List supported models
gateway	Run OpenAI-compatible OCR / VLM models on the VLM Run Gateway (alias gw)
config	Manage CLI configuration

Shell Completion

Install tab completion for your shell:

# Install completion
vlmrun --install-completion

# Show completion script (to customize)
vlmrun --show-completion