
Universal media extraction and analysis skill for Claude Code. Extracts structured knowledge from YouTube videos, web articles, local files, and pasted content.
概览
Universal media extraction and analysis skill for Claude Code. Extracts structured knowledge from YouTube videos, web articles, local files, and pasted content. - - transcript-based analysis OR visual analysis via Gemini - - fetch and analyze any URL - - video, audio, PDF, text files - - transcripts, meeting notes, any text - - multiple sources with comparison - - summary, golden nuggets, chapters, quotes, workflows, commands, breakdown, full extraction, visual analysis, download, batch - yt-dlp - YouTube data extraction - jq - JSON processing Visual analysis (Gemini multimodal) lets Claude see what's on screen in videos - code, terminal, UI, diagrams. It requires: - Python 3.8+ - The gemini-api-dev Claude Code skill - A Gemini API key Install the Gemini skill (pick one method): Then install the Python SDK (used by the extraction script): 1. Copy the skill to your Claude Code skills directory:
README
以下为原文 README 节选,完整内容请查看原文。
Media Extract
Universal media extraction and analysis skill for Claude Code. Extracts structured knowledge from YouTube videos, web articles, local files, and pasted content.
Features
- YouTube videos - transcript-based analysis OR visual analysis via Gemini
- Web articles - fetch and analyze any URL
- Local media - video, audio, PDF, text files
- Pasted content - transcripts, meeting notes, any text
- Batch mode - multiple sources with comparison
- 11 workflows - summary, golden nuggets, chapters, quotes, workflows, commands, breakdown, full extraction, visual analysis, download, batch
Prerequisites
Required
brew install yt-dlp jq
Required for visual analysis
Visual analysis (Gemini multimodal) lets Claude see what’s on screen in videos - code, terminal, UI, diagrams. It requires:
- Python 3.8+
- The gemini-api-dev Claude Code skill
- A Gemini API key
Install the Gemini skill (pick one method):
# Using Vercel Skills CLI
npx skills add google-gemini/gemini-skills --skill gemini-api-dev --global
# OR using Context7 Skills CLI
npx ctx7 skills install /google-gemini/gemini-skills gemini-api-dev
Then install the Python SDK (used by the extraction script):
pip install google-genai
Installation
- Copy the skill to your Claude Code skills directory:
cp -r media-extract ~/.claude/skills/content/media-extract
- (For visual analysis) Set up your Gemini API key:
cd ~/.claude/skills/content/media-extract
cp .env.example .env
# Edit .env and add your Gemini API key
- Make the fetch script executable:
chmod +x ~/.claude/skills/content/media-extract/fetch.sh
- Verify installation:
# Test yt-dlp
yt-dlp --version
# Test jq
jq --version
# Test Gemini skill + SDK
python3 -c "from google import genai; print('OK')"
Usage
The skill activates automatically when you share content with Claude Code:
# YouTube - transcript summary (default)
Summarize this: https://youtube.com/watch?v=...
# YouTube - visual analysis (sees code on screen)
Watch this and extract the code: https://youtube.com/watch?v=...
# YouTube - download
Download this video: https://youtube.com/watch?v=...
# Web article
Summarize this article: https://example.com/blog/post
# Local video
Analyze this video: /path/to/tutorial.mp4
# Pasted text
[paste transcript] Extract the golden nuggets from this
# Batch
Compare these two videos: [url1] [url2]
Available workflows
| Trigger | Workflow | Description |
|---|
推荐工具
换一个关键词,或者移除筛选条件。
安装
npx skillfish add alexadark/media-extract-skill