360 Secure SkillHub
360 安全技能中心

MiniMax 媒体生成 Unified MiniMax media generation skill for audio, image, and video creation with a single command entrypoint.MiniMax Skill 是一个统一的媒体生成技能，把文本转语音、文生图、文生视频三类能力收口为一个入口。安装后只需配置自己的 MINIMAX_API_KEY，即可通过统一命令生成音频、图片和视频，适合在 OpenClaw / Codex 类环境中直接调用。

该 Skill 通过 Python 统一命令调用 MiniMax API，支持文本转语音、文生图、图生图、文生视频和图生视频，并将生成的 MP3、JPEG 或 MP4 文件保存到指定位置；依赖 Python 3、requests 包和 MINIMAX_API_KEY 环境变量。

nmww
nmww
热度 72.31
最近更新 2026-05-11 17:41:50
分类：创作依赖：代码执行 / 读写文件 / 第三方 API应用场景：图像生成 / 视频编辑 / 故事创作 / 自动化脚本

安装技能

复制下面的话术发给你的 Agent，它会通过 secure-skills 检索并安装该技能。

执行命令安装此 skill： npx secure-skills-cli add minimax-skill
Metadata
name
minimax-skill
description
Unified MiniMax media generation skill for Token Plan workflows. Use when the user asks to generate audio, speech, TTS, narration, images, illustrations, posters, covers, thumbnails, videos, motion clips, text-to-video, image-to-video, or needs MiniMax API setup instructions across Linux, macOS, Windows, Docker, or CI.

MiniMax Media Generation

Use this skill to generate MiniMax audio, images, and videos from one command surface.

Requirements

  • Python 3
  • requests package
  • MINIMAX_API_KEY environment variable

For system-specific API key setup, read references/setup.md.

Unified Command

Run scripts/minimax.py with one mode: audio, image, or video.

python scripts/minimax.py {audio|image|video} [options]

Audio

Generate an MP3 file from text:

python scripts/minimax.py audio \
  --text "你好，今天我们聊一下 MiniMax。" \
  --output /tmp/minimax-audio.mp3

Useful options:

  • --model speech-2.8-turbo for lower latency
  • --model speech-2.8-hd for higher fidelity
  • --voice-id <voice_id> to select a voice when available

Image

Generate images into a directory:

python scripts/minimax.py image \
  --prompt "一只穿西装的橘猫，电影感，柔光" \
  --output /tmp/minimax-images \
  --aspect-ratio 1:1

Useful options:

  • --model image-01
  • --model image-01-live for stronger style control when available
  • --n 2 to request multiple images
  • --image-file <url_or_base64> for image-to-image workflows when supported

Outputs are saved as output-0.jpeg, output-1.jpeg, etc.

Video

Generate a video file:

python scripts/minimax.py video \
  --prompt "镜头缓慢推进，一只橘猫走过雨夜街道，霓虹灯反光" \
  --output /tmp/minimax-video.mp4

Useful options:

  • --model MiniMax-Hailuo-2.3
  • --first-frame-image <url_or_base64> for image-to-video
  • --last-frame-image <url_or_base64> for start/end-frame video
  • --subject-reference '<json>' for subject reference workflows
  • --timeout 1800 to control polling timeout in seconds

Direct Scripts

The unified command delegates to standalone scripts in the same skill:

  • scripts/generate_audio.py
  • scripts/generate_image.py
  • scripts/generate_video.py

Use these directly only when a narrower, mode-specific command is preferred.

Safe Sharing Notes

  • Do not include a real API key in the skill package.
  • Ask users to configure MINIMAX_API_KEY locally.
  • Keep generated media outside the skill folder, for example under /tmp or a project output directory.