A Claude Code Skill, joeseesun/anything-to-notebooklm, addresses the challenge of transforming diverse content sources into various usable formats like podcasts, presentations, mind maps, and quizzes. It targets technically savvy users, researchers, and engineers seeking an efficient way to process and repurpose information using natural language commands, integrating with NotebookLM for AI-driven generation.

How It Works

This project acts as a multi-source content processor, leveraging a Claude Code Skill to interpret natural language requests. It automatically identifies over 15 input types, including WeChat articles, web pages, YouTube videos, PDFs (with OCR for scanned documents), Office files, and ebooks. For specific sources like WeChat, it employs a custom "MCP" tool to bypass anti-scraping measures. The processed content is then uploaded to Google NotebookLM, which uses its AI capabilities to generate the desired output format, such as audio podcasts, PDF slides, or structured reports.

Quick Start & Requirements

• Primary install / run command: Installation involves cloning the repository into the ~/.claude/skills/ directory, running ./install.sh to install dependencies, configuring MCP, and restarting Claude Code. NotebookLM authentication is required via notebooklm login and notebooklm list.
• Non-default prerequisites and dependencies: Python 3.9+ and Git are required. The install.sh script handles other dependencies, including playwright install chromium for MCP.
• Links: GitHub Repository

Highlighted Details

• Supports over 15 content source formats, including WeChat articles (with anti-scraping bypass), YouTube videos (auto-subtitles), PDFs (scanned OCR), and audio (transcription).
• Generates multiple output formats: Podcasts, PPTs, Mind Maps, Quizzes, Videos, Reports, Infographics, and Flashcards, all triggered by natural language.
• Features "local-first" processing for sensitive content before uploading to NotebookLM.
• Handles batch processing for multiple files or ZIP archives.

Maintenance & Community

The project is maintained by Joe and primarily uses GitHub Issues and Discussions for community interaction. It shows signs of active development with recent commits and a growing number of stars and forks.

Licensing & Compatibility

The project is licensed under the MIT License, allowing for free use. However, commercial use is subject to the NotebookLM service terms and the copyright of the original content, with recommendations to use it primarily for personal learning and research.

Limitations & Caveats

Content length recommendations suggest 1000-10000 words for optimal results, with a maximum limit of approximately 500,000 words. While NotebookLM supports multiple languages, Chinese and English yield the best results. The generated podcast voice is from Google AI TTS.