video-commercial
Generate 30-second video commercials from a concept. Creates storyboard, generates scene images, adds narration via ElevenLabs, assembles final video. Use when asked to create commercials, promo videos, video ads, or short marketing videos.
Best use case
video-commercial is best used when you need a repeatable AI agent workflow instead of a one-off prompt.
Generate 30-second video commercials from a concept. Creates storyboard, generates scene images, adds narration via ElevenLabs, assembles final video. Use when asked to create commercials, promo videos, video ads, or short marketing videos.
Teams using video-commercial should expect a more consistent output, faster repeated execution, less prompt rewriting.
When to use this skill
- You want a reusable workflow that can be run more than once with consistent structure.
When not to use this skill
- You only need a quick one-off answer and do not need a reusable workflow.
- You cannot install or maintain the underlying files, dependencies, or repository context.
Installation
Claude Code / Cursor / Codex
Manual Installation
- Download SKILL.md from GitHub
- Place it in
.claude/skills/video-commercial/SKILL.mdinside your project - Restart your AI agent — it will auto-discover the skill
How video-commercial Compares
| Feature / Agent | video-commercial | Standard Approach |
|---|---|---|
| Platform Support | Not specified | Limited / Varies |
| Context Awareness | High | Baseline |
| Installation Complexity | Unknown | N/A |
Frequently Asked Questions
What does this skill do?
Generate 30-second video commercials from a concept. Creates storyboard, generates scene images, adds narration via ElevenLabs, assembles final video. Use when asked to create commercials, promo videos, video ads, or short marketing videos.
Where can I find the source code?
You can find the source code on GitHub using the link provided at the top of the page.
SKILL.md Source
# Video Commercial Generator
Create short video commercials (~30 seconds) from a concept through a multi-phase pipeline.
## Workflow
### Phase 1: Concept → Storyboard
Parse the concept into 5-6 scenes. Create `storyboard.json`:
```json
{
"title": "Project Name",
"concept": "Original concept text",
"duration": 30,
"scenes": [
{
"id": 1,
"duration": 5,
"visual": "Detailed image generation prompt",
"motion": "zoom-in|zoom-out|pan-left|pan-right|static",
"narration": "Voiceover text for this scene",
"overlay": "Optional text overlay"
}
]
}
```
### Phase 2: Generate Images
For each scene, generate an image using available image tools (nano-banana-pro or openai-image-gen).
Save to: `scenes/scene-{id}.png`
Save prompt metadata to: `scenes/scene-{id}.json`
### Phase 3: Generate Narration
Use ElevenLabs (sag skill) to generate voiceover from combined narration text.
Save to: `audio/narration.mp3`
### Phase 4: Create Video Clips
Use ffmpeg to create video clips from images with motion effects:
```bash
# Zoom in effect (5 seconds)
ffmpeg -loop 1 -i scene-01.png -vf "scale=8000:-1,zoompan=z='min(zoom+0.0015,1.5)':d=150:s=1920x1080" -t 5 -pix_fmt yuv420p clip-01.mp4
# Static with slight zoom
ffmpeg -loop 1 -i scene-01.png -vf "scale=2000:-1,zoompan=z='1.1':d=150:s=1920x1080" -t 5 -pix_fmt yuv420p clip-01.mp4
```
### Phase 5: Assemble Final Video
1. Concatenate clips
2. Add narration audio
3. Add background music (optional)
4. Add text overlays
5. Export final video
```bash
# Concat clips
ffmpeg -f concat -i clips.txt -c copy combined.mp4
# Add audio
ffmpeg -i combined.mp4 -i narration.mp3 -c:v copy -c:a aac -map 0:v -map 1:a final.mp4
```
## Project Structure
```
projects/video-commercial/{project-slug}/
├── concept.md # Original concept
├── storyboard.json # Scene breakdown
├── scenes/
│ ├── scene-01.png # Generated images
│ ├── scene-01.json # Prompt + metadata
│ └── ...
├── audio/
│ ├── narration.mp3 # ElevenLabs voiceover
│ └── music.mp3 # Background (optional)
├── clips/
│ ├── clip-01.mp4 # Individual scene videos
│ └── ...
├── output/
│ └── final.mp4 # Assembled video
└── review.md # Review notes + iterations
```
## Review Process
After each phase, update `review.md` with:
- What was generated
- User feedback
- Scenes to regenerate
- Approved scenes (mark with ✓)
Allow regeneration of individual scenes without redoing entire project.
## Commands Reference
| Command | Action |
|---------|--------|
| `new commercial "{concept}"` | Create project + storyboard |
| `show storyboard` | Display current storyboard |
| `generate images` | Generate all scene images |
| `regen scene {n}` | Regenerate specific scene |
| `preview` | Show all scenes with status |
| `narrate` | Generate voiceover |
| `render` | Assemble final video |
| `status` | Show project progress |
## Notes
- Projects stored in workspace: `projects/video-commercial/`
- Image aspect ratio: 16:9 (1920x1080)
- Default scene duration: 5 seconds
- Narration voice: Configure in project or use defaultRelated Skills
vidu-video
使用 Vidu Q3 Pro 模型生成视频。当用户想要文生视频、生成带音频的视频,或提到 vidu 时使用此 skill。
videodb-skills
Upload, stream, search, edit, transcribe, and generate AI video and audio using the VideoDB SDK.
videocut:安装
环境准备。安装依赖、下载模型。触发词:安装、环境准备、初始化
video
Generate videos using fal.ai (Wan, Kling) or Sora. Text-to-video and image-to-video.
video-toolkit
Intelligent video processor for downloading media and extracting transcripts from YouTube and 1000+ supported sites. Automatically handles format selection, subtitle extraction, and post-processing.
video-processing-editing
FFmpeg automation for cutting, trimming, concatenating videos. Audio mixing, timeline editing, transitions, effects. Export optimization for YouTube, social media. Subtitle handling, color grading, batch processing. Use for videogen projects, content creation, automated video production. Activate on "video editing", "FFmpeg", "trim video", "concatenate", "transitions", "export optimization". NOT for real-time video editing UI, 3D compositing, or motion graphics.
video-analyzer
鏅鸿兘鍒嗘瀽 Bilibili/YouTube/鏈湴瑙嗛锛岀敓鎴愯浆鍐欍€佽瘎浼板拰鎬荤粨銆傛敮鎸佸叧閿抚鎴浘鑷姩宓屽叆銆?
Media Uploader - R2/S3 with video download
Upload files or download videos from popular platforms (YouTube, Vimeo, Bilibili, etc.) and upload to Cloudflare R2, AWS S3, or any S3-compatible storage with secure presigned download links.
ltxv2-video
Build LTX-V2 19B video workflows — text-to-video, image-to-video, distilled model, camera control LoRAs, and two-stage upscaling
edu-video-analyzer
Analyze educational YouTube channels for classroom adoption potential, curriculum alignment, and pedagogical effectiveness. Use when comparing educational video content (like MRU vs Crash Course), evaluating teaching methodologies, identifying content gaps for course design, or developing educational video strategy focused on student learning outcomes rather than monetization.
Automate YouTube Top-Ten Video Creation with OpenAI and Safe Image Search
Integrates OpenAI API for content generation, Bing Image Search API for safe image retrieval, and Pexels API for video footage. Handles authentication via Bearer token, enforces safe search, formats ChatGPT responses into a top-ten list, and includes error handling for API failures.
apex-video-generator
Generate real estate marketing videos from property data. Use when creating property showcases, social media content, market reports, or neighborhood tours. Integrates Firecrawl scraped data with Remotion rendering.