# Video I2i

> Use when user needs to transform or edit images using AI. Independent image-to-image command for converting reference images to different styles or content.

- Skill: `t0ugh/video-i2i` (Agent Skill)
- Install (CLI): `npx skillmds@latest add t0ugh/video-i2i`
- Raw SKILL.md: https://api.skillmd.com/api/skills/t0ugh/video-i2i/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: T0UGH (https://skillmd.com/u/t0ugh)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/t0ugh/video-i2i

---


# video-i2i - 图生图

## Overview

Transform or edit images using AI. Independent command for converting reference images to different styles or content.

## Usage

> **注意**：如果是首次使用，确保已安装 videoclaw：`uvx videoclaw --help`

```bash
# Transform image
videoclaw i2i -i input.png -p "转换为九宫格角色图" -o output.png

# Specify provider
videoclaw i2i -i input.png -p "描述" -o output.png --provider volcengine

# Specify model
videoclaw i2i -i input.png -p "描述" -o output.png --provider gemini --model gemini-3-pro-image-preview
```

## Parameters

- `--input, -i`: Input image path (required)
- `--prompt, -p`: Transformation prompt (required)
- `--output, -o`: Output file path (required)
- `--provider`: Model provider (optional, default: volcengine)
- `--model`: Model name (optional)

## Supported Providers

- `volcengine` - Volcano Engine Seedream (default)
- `dashscope` - Alibaba Wan2.6
- `gemini` - Google Gemini
- `mock` - For testing

## Examples

```
Claude Code: videoclaw i2i -i reference.jpg -p "转换为九宫格角色图" -o character_grid.png
```