---
title: "Introduction"
description: "Generate text, images, video, and audio from the terminal."
canonical_url: "https://ai-cli.dev/docs"
---
# Introduction

ai-cli is a tiny, agent-native CLI for generating text, images, video, and audio with dead-simple commands. It uses the [Vercel AI SDK](https://sdk.vercel.ai) and [AI Gateway](https://vercel.com/docs/ai-gateway) for unified access to hundreds of models.

## What it does

- **Image generation**: Create images from text or existing images.

- **Video generation**: Generate video from text or images.

- **Text generation**: Generate text from prompts with any language model.

- **Audio generation**: Generate speech or transcribe audio files and streams.

- **Multi-model comparison**: Run the same prompt across multiple models in parallel.

- **Stdin piping**: Pipe any content in and out for composable workflows.

- **Inline preview**: See generated images, video frames, and speech playback directly in your terminal.

## Quick start

1. **Install globally**

   ```bash
   npm install -g ai-cli
   ```

2. **Set your API key**

   ```bash
   export AI_GATEWAY_API_KEY="your-key"
   ```

3. **Generate something**

   ```bash
   ai image "a cute dog"
   ai video "a spinning triangle"
   ai text "explain quantum computing"
   ai audio speak "Thanks for trying ai-cli"
   ai audio transcribe recording.mp3
   ```

## Commands

- [**image**](/docs/commands#image): Generate images from a prompt.

- [**video**](/docs/commands#video): Generate video from a prompt or image.

- [**text**](/docs/commands#text): Generate text from a prompt.

- [**audio**](/docs/commands#audio): Generate speech or transcribe audio.

- [**models**](/docs/commands#models): List available models from the gateway.

> **Tip**:
>
> ai-cli is designed to be used by both humans and agents. Predictable output behavior, JSON metadata mode, and stdin/stdout piping make it easy to integrate into any workflow.