# Image to Video AI MCP Server

A lightweight Model Context Protocol server for planning image-to-video shots, building production-ready motion prompts, and preparing reliable generation workflows.

## Installation

Run directly with npx:

```bash
npx -y mcp-imagetovideoai-server
```

Or install globally:

```bash
npm install -g mcp-imagetovideoai-server
```

## MCP Configuration

```json
{
  "mcpServers": {
    "image-to-video-ai": {
      "command": "npx",
      "args": ["-y", "mcp-imagetovideoai-server"]
    }
  }
}
```

## Tools

### `image_to_video_build_prompt`

Builds a structured motion prompt from the subject, primary motion, environment, camera, lighting, timing, style, and artifact constraints.

Required inputs:

- `subject`
- `motion`

### `image_to_video_plan`

Creates a concise shot plan with motion hierarchy, camera guidance, continuity checks, optional end-frame control, and an iteration strategy.

Required inputs:

- `goal`
- `image_description`

### `image_to_video_get_started`

Returns a localized workspace URL and a short preflight checklist. Supported locales are English (`en`), Japanese (`ja`), and Chinese (`zh`).

## Example

Ask your MCP client:

```text
Build a five-second prompt for a product photo. The bottle should rotate slowly,
the camera should make a subtle dolly-in, and the label must remain stable.
```

Use the resulting plan with the [Image to Video AI generator](https://imagetovideoai.pro/) to create and iterate on the clip.

## Scope

This package prepares prompts and shot plans. It does not request account credentials, store API keys, upload source media, or claim to call a private generation API.

## License

MIT