Skip to main content
Glama

🎯 Anything → NotebookLM

Multi-source Content Intelligent Processor: Anything → Podcast / PPT / Mind Map / Quiz

License: MIT Python 3.9+ PRs Welcome GitHub stars GitHub forks GitHub issues GitHub last commit

Quick StartSupported FormatsUsage ExamplesPaywall BypassFAQ


✨ What is this?

A Claude Code Skill that turns anything into any format using natural language.

你说:把这篇微信文章生成播客
AI :✅ 8 分钟播客已生成 → podcast.mp3

你说:这个付费文章做成思维导图
AI :✅ 自动绕过付费墙 → 思维导图已生成

你说:这期播客(小宇宙)做成 PPT
AI :✅ 自动转录音频 → 25 页 PPT 已生成

Core Capabilities: Multi-source content acquisition (including paywall bypass) → Upload to Google NotebookLM → AI-generated target format.


Related MCP server: KHU Notebook Research Assistant

🚀 Supported Content Sources (15+)

📱 Social & Media

  • WeChat Official Accounts (MCP browser simulation)

  • X/Twitter (Tweets + long threads)

  • YouTube Videos (Automatic subtitle extraction)

  • Podcasts (Xiaoyuzhou / Ximalaya / Bilibili)

🌐 Web (Including Paywall Bypass)

  • 300+ Paywalled Sites (NYT/WSJ/FT/Economist...)

  • Any Public Webpage (News, blogs, documents)

  • Search Keywords (Automatic result aggregation)

📚 E-books & Documents

  • PDF (Supports OCR for scanned copies)

  • EPUB E-books

  • Markdown (.md)

  • Plain Text (.txt)

📄 Office Documents

  • Word (.docx)

  • PowerPoint (.pptx)

  • Excel (.xlsx)

🖼️ Others

  • Images (JPEG/PNG, automatic OCR)

  • Audio (WAV/MP3, automatic transcription)

  • ZIP Archives (Batch processing)


🛡️ Paywall Bypass

Core Feature: Automatically detects and bypasses paywalls for 300+ news websites.

Bypass Strategy (6-Layer Cascade)

Level 1: 代理服务(r.jina.ai / defuddle.md)
    ↓ 失败
Level 2: 站点专属 Bot UA(Googlebot ~50站 / Bingbot ~4站)
    ↓ 失败
Level 3: 通用绕过(UA伪装 + X-Forwarded-For + Referer伪装 + AMP + EU IP)
    ↓ 失败
Level 4: archive.today 存档(CAPTCHA 自动检测)
    ↓ 失败
Level 5: Google Cache
    ↓ 失败
Level 6: agent-fetch 本地工具

Supported Paywalled Sites (Partial)

Category

Sites

🇺🇸 US Media

NYT, WSJ, Bloomberg, Washington Post, The Information, Forbes, WIRED, The New Yorker, The Atlantic, USA Today, Boston Globe, LA Times, Chicago Tribune, Seattle Times, MIT Tech Review, Foreign Affairs

🇬🇧 UK Media

FT, The Times, The Telegraph, The Economist

🇩🇪 German Media

Spiegel, Zeit, Sueddeutsche, FAZ, Handelsblatt

🇫🇷 French Media

Le Monde, Le Figaro, Le Parisien

🇦🇺 Australian Media

The Australian, SMH, The Age, Brisbane Times

🇨🇳 Chinese Media

SCMP, Medium

🌐 Others

Haaretz, NZ Herald, Statista, Quora

Bypass Technology (Learned from Bypass Paywalls Clean)

Technology

Principle

Coverage

Googlebot UA + X-Forwarded-For

Search engine crawler whitelist, direct full-text access

~50 sites

Bingbot UA

Same as above, some sites are more friendly to Bing

~4 sites

Cookie Clearing + Referer Spoofing

Clear metered cookies, spoof origin from Google/Facebook/Twitter

Metered paywalls

AMP Pages

AMP versions have weaker paywall implementations

~10 sites

JSON-LD Extraction

Extract articleBody from HTML-embedded structured data

Universal

archive.today

Retrieve saved content from web archives

Fallback solution


🎨 What can be generated?

Output Format

Purpose

Trigger Word Examples

🎙️ Podcast

Listen during commute

"Generate podcast", "Make into audio"

📊 PPT

Team sharing

"Make into PPT", "Generate slides"

🗺️ Mind Map

Clarify structure

"Draw a mind map", "Generate mind map"

📝 Quiz

Self-assessment

"Generate Quiz", "Create questions"

🎬 Video

Visualization

"Make a video"

📄 Report

Deep analysis

"Generate report", "Write a summary"

📈 Infographic

Data visualization

"Make an infographic"

📋 Flashcards

Memory consolidation

"Make into flashcards"


⚡ Quick Start

Prerequisites

  • ✅ Python 3.9+

  • ✅ Git (Pre-installed on macOS/Linux)

That's it! Other dependencies are installed automatically with one click.

Installation (3 Steps)

# 1. 克隆到 Claude skills 目录
cd ~/.claude/skills/
git clone https://github.com/joeseesun/qiaomu-anything-to-notebooklm
cd qiaomu-anything-to-notebooklm

# 2. 一键安装所有依赖
./install.sh

# 3. 按提示配置 MCP,然后重启 Claude Code

First Use

# NotebookLM 认证(只需一次)
notebooklm login
notebooklm list  # 验证成功

# 环境检查(可选)
./check_env.py

Podcast Transcription Configuration (Optional)

To use the Xiaoyuzhou/Ximalaya/Bilibili transcription feature, configure the GetNote API:

export GETNOTE_API_KEY="your_api_key"
export GETNOTE_CLIENT_ID="your_client_id"

💡 Usage Examples

Scenario 1: Paywalled Article → Podcast

你:把这篇 The Information 文章生成播客 https://www.theinformation.com/articles/...

AI 自动执行:
  ✓ 检测付费墙 → Googlebot UA 绕过
  ✓ 获取完整文章内容
  ✓ 上传到 NotebookLM
  ✓ 生成播客

✅ 结果:/tmp/article_podcast.mp3

Scenario 2: Podcast (Xiaoyuzhou) → PPT

你:这期小宇宙播客做成 PPT https://xiaoyuzhoufm.com/episode/...

AI 自动执行:
  ✓ Get笔记 API 转写音频(2-5 分钟)
  ✓ 上传转写文本到 NotebookLM
  ✓ 生成 PPT

✅ 结果:/tmp/podcast_slides.pdf(25 页)

Scenario 3: E-book → Deep Analysis

你:深度分析这本书 /Users/joe/Books/sapiens.epub

AI 自动执行:
  ✓ 提取 EPUB 全文
  ✓ 上传到 NotebookLM
  ✓ 生成 12 个问题(3 轮递进:概览→深度挖掘→综合反刍)
  ✓ 逐轮提问,后轮受益于前轮对话上下文
  ✓ 输出结构化 JSON

✅ 结果:/tmp/sapiens_analysis.json(12 个问答,含核心观点、论证拆解、矛盾分析、认知改变)

Scenario 4: X/Twitter Thread → Mind Map

你:这个推文线程做成思维导图 https://x.com/user/status/123...

AI 自动执行:
  ✓ 代理级联获取推文内容(含完整线程)
  ✓ 上传到 NotebookLM
  ✓ 生成思维导图

✅ 结果:/tmp/tweet_mindmap.json

Scenario 5: WeChat Article → Lark/Feishu Doc (Deep Analysis)

你:深度分析这篇微信文章并写入飞书 https://mp.weixin.qq.com/s/abc123

AI 自动执行:
  ✓ MCP 浏览器模拟抓取微信文章
  ✓ 上传到 NotebookLM
  ✓ 生成 10 个问题并递归提问
  ✓ 格式化为飞书 Markdown
  ✓ 自动创建飞书文档

✅ 结果:飞书文档已创建(含完整问答)

🎯 Core Features

🧠 Intelligent Recognition

Automatically determines input type, no manual specification needed.

https://mp.weixin.qq.com/s/xxx        → 微信公众号
https://xiaoyuzhoufm.com/episode/xxx  → 小宇宙播客
https://x.com/user/status/xxx         → X/Twitter
https://youtube.com/watch?v=xxx       → YouTube 视频
/path/to/file.epub                    → EPUB 电子书
"搜索 'AI 趋势'"                       → 搜索查询

🛡️ Automatic Paywall Bypass

No manual handling required, automatically detected and bypassed.

检测付费墙 → 选择最佳策略 → 获取完整内容
     ︿________全自动________︿

🚀 Fully Automated Processing

From acquisition to generation, all in one go.

输入 → 获取 → 转换 → 上传 → 生成 → 下载
      ︿___________全自动___________︿

🌐 Multi-source Integration

Supports mixing multiple content sources.

付费文章 + YouTube 视频 + EPUB + 播客 → 综合报告

📦 Technical Architecture

┌──────────────────────────────────────────┐
│            用户自然语言输入                │
│  "把这个付费文章生成播客 https://..."     │
└──────────────────┬───────────────────────┘
                   │
                   ▼
┌──────────────────────────────────────────┐
│         Claude Code Skill                 │
│  • 智能识别内容源类型                      │
│  • 自动调用对应工具                        │
└──────────┬───────────────────────────────┘
           │
   ┌───────┴───────┐
   │               │
   ▼               ▼
┌──────────┐  ┌──────────────┐  ┌──────────┐  ┌──────────┐
│ 微信 MCP  │  │ 付费墙绕过   │  │ 播客转写  │  │ markitdown│
│ 浏览器模拟 │  │ 6层级联策略  │  │ Get笔记API│  │ 文件转换  │
└─────┬────┘  └──────┬───────┘  └─────┬────┘  └─────┬────┘
      │              │                 │              │
      └──────────────┴─────────────────┴──────────────┘
                           │
                           ▼
              ┌────────────────────────┐
              │    NotebookLM API      │
              │  • 上传内容源           │
              │  • AI 生成目标格式      │
              └───────────┬────────────┘
                          │
                          ▼
              ┌────────────────────────┐
              │       生成的文件        │
              │ .mp3 / .pdf / .json    │
              └────────────────────────┘

📂 Project Structure

qiaomu-anything-to-notebooklm/
├── SKILL.md                          # Skill 定义文件
├── README.md                         # 本文件
├── main.py                           # 主入口:CLI 智能处理器
├── install.sh                        # 一键安装脚本
├── check_env.py                      # 13 项环境检查
├── package.sh                        # 打包分享脚本
├── requirements.txt                  # Python 依赖
├── LICENSE                           # MIT
├── scripts/
│   ├── fetch_url.sh                  # URL 抓取 + 付费墙绕过(6 层级联)
│   └── get_podcast_transcript.py     # 播客/视频转写(Get笔记 API)
├── wexin-read-mcp/                   # 微信公众号 MCP 服务器
│   └── src/
│       ├── server.py                 # MCP 入口
│       ├── scraper.py                # Playwright 浏览器模拟
│       └── parser.py                 # HTML 解析
└── feishu-read-mcp/                  # 飞书文档 MCP 服务器
    └── src/
        ├── server.py                 # MCP 入口
        ├── scraper.py                # 飞书文档抓取
        ├── parser.py                 # HTML → Markdown
        └── image_handler.py          # 图片处理

🔧 Advanced Usage

Deep Analysis Mode

python main.py https://example.com/article --deep-analysis
# 自动生成 12 个问题(3 轮递进:概览→深度挖掘→综合反刍),逐轮提问,输出结构化 JSON

Three-Round Progressive Strategy:

Round

Number of Questions

Purpose

Examples

Round 1: Overview & Framework

4

Establish overall understanding

Summarize topic, list structure, extract core arguments, uncover disruptive content

Round 2: Deep Dive

5

Explore details

Deconstruct logic, analyze contradictions, distill core insights, propose sharp criticism

Round 3: Synthesis & Reflection

3

Cognitive upgrade

Biggest cognitive shift, action guide, recommendation reasons

NotebookLM maintains context within the same session; subsequent round questions automatically benefit from previous answers, forming a true "progressive" deep analysis.

Lark/Feishu Doc Output

python main.py ./book.epub --deep-analysis --to-feishu
# 深度分析后自动创建飞书文档

Batch Processing

把这些文章都生成播客:
1. https://mp.weixin.qq.com/s/abc123
2. https://www.wsj.com/articles/...
3. /Users/joe/notes.md

🐛 Troubleshooting

MCP Tool Not Found

python ~/.claude/skills/qiaomu-anything-to-notebooklm/wexin-read-mcp/src/server.py
cd ~/.claude/skills/qiaomu-anything-to-notebooklm/wexin-read-mcp
pip install -r requirements.txt
playwright install chromium

NotebookLM Authentication Failed

notebooklm login     # 重新登录
notebooklm list      # 验证

Paywall Bypass Failed

Some hard-paywall sites (like The Information) do not send content server-side, requiring archive.today. The script will automatically detect and prompt:

⚠️  archive.ph needs human verification.
   已自动打开浏览器,请完成验证后重试

Environment Check

./check_env.py       # 13 项全面检查
./install.sh         # 重新安装

❓ FAQ

A: NotebookLM supports multiple languages; Chinese and English perform best.

A: Google AI voice synthesis. English is a dialogue between two AI hosts; Chinese is a single-person narration.

A: This tool is for personal study and research only. The technical principle is based on search engine whitelists (Googlebot/Bingbot) and does not crack any encryption. It is recommended to support high-quality news media by purchasing subscriptions.

A:

  • Minimum: ~500 words

  • Maximum: ~500,000 words

  • Recommended: 1,000-10,000 words for best results

A: WeChat Official Accounts have anti-scraping measures; MCP uses Playwright browser simulation to bypass them. Other content sources (webpages, YouTube, PDF) do not require MCP.

A: Supported via GetNote API for Xiaoyuzhou, Ximalaya, and Bilibili videos. YouTube is handled directly by NotebookLM.


🙏 Acknowledgments

📄 License

MIT License - For personal study and research use only


If you find this useful, please give it a ⭐ Star!

Made with ❤️ by Joe · Twitter @vista8 · WeChat Official Account "向阳乔木推荐看"

A
license - permissive license
Not graded
quality - not tested
D
maintenance

Maintenance

Maintainers
Response time
0dRelease cycle
2Releases (12mo)
Commit activity
Issues opened vs closed

Resources

Unclaimed servers have limited discoverability.

Looking for Admin?

If you are the server author, to access and configure the admin panel.

Related MCP Servers

  • A
    license
    Not graded
    quality
    D
    maintenance
    An MCP server that uses the Docling toolkit to convert various document formats, including PDFs, Office files, images, and audio, into clean Markdown for AI processing. It supports multiple processing pipelines like VLM and ASR with intelligent auto-detection and job queue management.
    2
    MIT
  • F
    license
    B
    quality
    D
    maintenance
    An MCP server that interfaces with Google NotebookLM to enable autonomous academic research and systematic knowledge management. It allows users to perform deep web searches and automatically generate study artifacts like research reports, presentation slides, and audio overviews.
    6
  • A
    license
    A
    quality
    Not graded
    maintenance
    An MCP server that enables AI agents to instantly convert Markdown into beautiful, ready-to-deliver Word, PDF, HTML with sidebars, and Slideshow documents, bridging the "last mile" of AI content generation.
    7
    28
    1

View all related MCP servers

Related MCP Connectors

  • OCR, transcription, file extraction, and image generation for AI agents via MCP.

  • MCP server for Google Veo AI video generation

  • MCP server for Wan AI video generation

View all MCP Connectors

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/joeseesun/qiaomu-anything-to-notebooklm'

If you have feedback or need assistance with the MCP directory API, please join our Discord server