ad_video_create

Generates 10-second advertisement videos from product images using image composition and video diffusion.

1.2k|126|Updated Mar 14, 2025
One-click install
npx skills add https://github.com/inclusionAI/AWorld --skill ad-video-create
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: ad_video_create
Source: https://github.com/inclusionAI/AWorld/tree/main/aworld-skills/ad_video_create_skill
Command: npx skills add https://github.com/inclusionAI/AWorld --skill ad-video-create

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires Pillow, ffmpeg.

What problem does it solve?

Creating commercial-quality product advertisement videos normally requires filming, editing, and audio production expertise. This Skill automates the entire pipeline, turning a single product image into a finished 10-second ad video with environment, character, motion, and sound.

Core Features & Use Cases

  • Automated Asset Pipeline: Analyzes product images, optionally generates a matching character, composes both into a realistic lifestyle scene, and converts the result into a dynamic video.
  • Commercial Shot Language: Video prompts follow advertising conventions—visual hooks, hero product shots, detail showcases, function expression, and dynamic camera motion.
  • Flexible Audio Handling: Merges user-provided MP3 files via FFmpeg or generates AI background music when no audio is supplied.
  • Use Case: An e-commerce seller uploads a cat tower photo and a cat photo; the Skill composes them into a cozy living room scene and outputs a 720p, 10-second ad video with background music ready for social media.

Quick Start

Create an advertisement video from the product image in this directory, adding a suitable character and background music.

Frequently Asked Questions about ad_video_create

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I create a product advertisement video from images?

Provide a product image and optionally a character image and MP3 audio file. The workflow analyzes the product, composes it with a character into a realistic environment, then generates a 10-second video using video diffusion with commercial shot language.

What inputs are required to generate an ad video?

Only a product image is required. A character or subject image and an MP3 audio file are optional; missing characters are generated automatically, and missing audio is replaced with AI-generated background music.

Can I use my own background music in the generated video?

Yes, place an MP3 file in the working directory. The video is first generated without sound, then FFmpeg merges your audio using AAC encoding at 192kbps, trimmed to the 10-second video length.

Why do multiple characters appear in my composed image?

Image generation models may add extra characters without explicit constraints. Add a directive like "ONLY ONE character, no other characters" to the composition prompt to enforce the single-character requirement.

What video resolution and duration does the workflow produce?

The workflow outputs 720p video at 24 fps with a 10-second duration, optimized for social media and e-commerce platforms. These parameters balance visual quality against file size and generation time.

What tools are needed to run the ad video workflow?

The workflow requires Python 3.8+ with PIL/Pillow for image compression, FFmpeg for audio-video merging, and access to media comprehension, image generation, and video diffusion AI services.