YouTube Lecture Notes MCP
Automatically extracts subtitles, captures screenshots at specified timestamps, and generates structured Obsidian markdown notes from YouTube videos using Claude AI.
README
<div align="center">
📝 YouTube Lecture Notes MCP
<img src="assets/youtube-logo.png" width="80" alt="YouTube" /> <img src="assets/claude-logo.png" width="80" alt="Claude AI" /> <img src="assets/obsidian-logo.png" width="80" alt="Obsidian" />
Automatically generate professional study notes from YouTube videos
YouTube + Claude AI + Obsidian = Supercharged Learning
Never manually take screenshots and paste them into notes again! This MCP server integrates with Claude Desktop to automatically extract subtitles, capture screenshots at key timestamps, and generate beautifully formatted Obsidian notes from any YouTube tutorial or lecture.
✨ Features
- Auto-extract subtitles from any YouTube video with captions
- Capture screenshots at timestamps you specify
- Generate structured notes in Obsidian markdown format
- Organized metadata with tags, dates, and video links
- Full transcripts included with timestamps
- Preview subtitles before processing
- Customizable paths for your Obsidian vault organization
- Works with Claude AI for natural language interaction
🎯 Perfect For
Learn from YouTube videos on any topic:
| Category | Examples |
|---|---|
| Programming | Python, JavaScript, React, Web Development |
| Creative Skills | Photoshop, Premiere Pro, Blender, Video Editing |
| Business & Productivity | Excel, Data Analysis, Project Management |
| Academic | Math, Physics, Chemistry, Biology |
| IT & Networking | CCNA, AWS, CompTIA, Cybersecurity |
| Personal Development | Languages, Fitness, Cooking, Music |
🚀 Quick Start
Prerequisites
Before installation, make sure you have:
- Python 3.10+ - Download here
- ⚠️ Check "Add Python to PATH" during installation
- FFmpeg - For capturing video screenshots
- Claude Desktop - Get it here
- Obsidian (optional but recommended) - Download
Installation
Option 1: Automated Install (Recommended)
- Clone or download this repository
- Open PowerShell in the project folder
- Run the installer:
.\install.ps1
The script will:
- ✅ Check all prerequisites
- ✅ Install required Python packages
- ✅ Set up directory structure
- ✅ Generate Claude Desktop configuration
Option 2: Manual Install
See the detailed SETUP_GUIDE.md for step-by-step instructions.
💡 Usage
Once installed, just talk to Claude naturally in Claude Desktop!
Example 1: Process a Python Tutorial
Hey Claude, process this Python tutorial:
https://www.youtube.com/watch?v=rfscVS0vtbw
Capture screenshots at: 2:30, 8:15, 15:00, 22:45, 35:00
This is Day 1 of my learning journey.
Example 2: Preview Before Processing
Can you show me the subtitles for this video first?
https://www.youtube.com/watch?v=example
Then I'll tell you which timestamps to capture.
Example 3: Custom Organization
Configure my YouTube notes:
- Vault: D:\Learning\Obsidian
- Notes folder: Courses/Web Development
- Images folder: Courses/Web Development/assets
What You Get
Each processed video creates:
1. Markdown Note File
Day 01 - Complete Python Course for Beginners.md
2. Screenshot Images
day1_abc123_1_150s.jpg
day1_abc123_2_495s.jpg
day1_abc123_3_900s.jpg
3. Structured Content
---
date: 2026-01-16
day: 1
tags: [lecture, video-notes, learning]
video: "https://youtube.com/watch?v=..."
---
# Day 1: Complete Python Course for Beginners
> **Source**: [freeCodeCamp.org](https://youtube.com/...) > **Duration**: 4:26:54
> **Study Date**: 2026-01-16
## Summary
[Automatic summary from transcript]
## Key Concepts & Notes
### Timestamp 02:30

**Context from lecture:**
> In this section, we cover variables and data types...
**Key Points:**
- [Your notes here]
## Key Terms & Definitions
| Term | Definition |
| ---- | ---------- |
| | |
## Review Questions
1.
2.
## Personal Notes & Observations
[Your thoughts]
## Full Transcript
[Complete timestamped transcript]
🛠️ Available Tools
The MCP server provides three tools accessible through Claude:
1. process_youtube_lecture
Main tool to process a video:
- Extracts subtitles
- Captures screenshots at your timestamps
- Generates complete Obsidian note
Parameters:
url- YouTube video URL (required)timestamps- Comma-separated times like "5:30, 12:00, 25:00" (required)day_number- Your study day number (required)vault_path- Custom vault path (optional)notes_folder- Custom notes folder (optional)images_folder- Custom images folder (optional)
2. get_video_subtitles
Preview the full transcript before processing:
- Returns complete subtitles with timestamps
- Helps you identify key moments
Parameters:
url- YouTube video URL (required)
3. configure_paths
Set or view your default paths:
- Configure Obsidian vault location
- Set custom folder structure
Parameters:
vault_path- New vault path (optional)notes_folder- New notes folder (optional)images_folder- New images folder (optional)
📂 File Structure
After installation, your files will be organized like this:
C:\Users\YourName\
├── mcp-servers\
│ └── youtube-lecture-notes\
│ ├── server.py
│ └── claude_config_snippet.json
│
└── Documents\
└── ObsidianVault\
└── Lecture Notes\
├── Day 01 - Python Basics.md
├── Day 02 - Advanced Functions.md
└── images\
├── day1_xxx_1_150s.jpg
├── day1_xxx_2_495s.jpg
└── day2_xxx_1_300s.jpg
🔧 Troubleshooting
Common Issues
"python is not recognized"
- Reinstall Python and check "Add to PATH"
- Or specify full path in config:
C:\\Python312\\python.exe
"ffmpeg is not recognized"
- Install FFmpeg:
winget install ffmpeg - Restart terminal after installation
- Verify:
ffmpeg -version
"No subtitles found"
- Check if the video has captions enabled on YouTube
- Try with a different video to test
- Some videos may not have subtitles available
MCP not connecting
- Verify paths in
claude_desktop_config.jsonuse double backslashes\\ - Restart Claude Desktop completely
- Check Claude Desktop logs in Settings > Developer
Screenshots not capturing
- Ensure FFmpeg is installed correctly
- Video may be region-restricted
- Some videos block downloading
For more help, see the SETUP_GUIDE.md or open an issue on GitHub.
🎓 Tips for Best Results
- Preview first - Use
get_video_subtitlesto scan content before choosing timestamps - Pick key moments - Choose 4-6 important timestamps per video (don't overdo it!)
- Watch at 2x speed - Quickly identify where major concepts are introduced
- Fill in blanks - The notes provide structure, but add your own insights
- Organize with tags - Customize the
tags:in frontmatter for your topics - Review regularly - Use the review questions section to test yourself
🎨 Customize Your Note Style
The generated notes are fully customizable to match your personal learning style and preferences!
What You Can Customize
1. Note Structure & Sections
- Add custom sections for your needs (code snippets, formulas, workflows)
- Reorder sections to match your thinking process
- Remove sections you don't use
- Create templates for different topics
2. Formatting & Style
- Change heading styles and hierarchy
- Customize emoji usage and visual markers
- Adjust table layouts for terms and definitions
- Modify YAML frontmatter tags and metadata
3. Organization
- Custom folder structures by topic, difficulty, or course
- Tag systems for easy filtering and linking
- Naming conventions that work for you
- Integration with your existing Obsidian vault structure
4. Content Enhancement
- Add your own code blocks with syntax highlighting
- Include links to external resources
- Embed additional images or diagrams
- Create internal wiki-links to related notes
How to Customize
See the comprehensive CUSTOMIZATION.md guide for:
- Step-by-step customization instructions
- Pre-made templates for different learning styles
- Topic-specific note structures (programming, design, business, etc.)
- Advanced Obsidian integration tips
- Examples of customized notes
Quick customization tips:
- Edit
server.pylines 254-349 to change the default note template - Modify tags in the frontmatter after generation
- Use Obsidian templates for post-processing
- Create Obsidian CSS snippets for custom styling
💡 Your notes, your way - Start with the default structure and evolve it as you discover what works best for your learning!
🤝 Contributing
Contributions are welcome! Feel free to:
- 🐛 Report bugs
- 💡 Suggest new features
- 🔧 Submit pull requests
- 📖 Improve documentation
Acknowledgments
Built with:
📚 Documentation
- SETUP_GUIDE.md - Detailed installation and setup instructions
- CLAUDE.md - Instructions for Claude AI on how to use this MCP effectively
- EXAMPLE_NOTE.md - Sample note structure and template
- CUSTOMIZATION.md - Complete guide to customizing your note style
📬 Support & Community
- Documentation: Check SETUP_GUIDE.md
- For Claude AI: See CLAUDE.md for usage guidelines
- Issues: Report bugs here
- Discussions: Share your use cases and tips
- License: MIT License
<div align="center">
Made with ❤️ for learners everywhere by Mohsin Khadim
</div>
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。