Anything to NotebookLM
github.com/joeseesun/qiaomu-anything-to-notebooklm- Category
- AI Agents
- Rank
- No. 1667Tools index
Previous survey · No. 1561 ·
- Type
- APP
- Builder
- joeseesun
- GitHub
- 6.0k stars
- Latest release
- v1.0.1
- Date
About
A Claude skill that turns multi-source content into NotebookLM-ready inputs.
What it does
Give it a web link, media URL, search phrase, or local document. It identifies the source, extracts or transcribes the material, sends it to NotebookLM, then can request podcasts, slide decks, mind maps, quizzes, reports, videos, infographics, or flashcards. An optional analysis mode asks progressively deeper questions and saves structured answers.
Why it's ranked here
The breadth is compelling, but the implementation looks uneven. It combines document conversion, browser-based extraction, podcast transcription, NotebookLM generation, and staged questioning behind natural-language requests. However, several workflows depend on external services, authentication, manual configuration, or access-bypass tactics. The supplied code also exposes a likely MCP startup error, which weakens confidence in immediate use.
What's good
Input detection covers common URLs and document types without requiring users to choose a parser. EPUB extraction preserves a simple text pipeline, while podcast links can become transcripts through the Get笔记 service. The analysis workflow asks twelve questions across overview, detailed criticism, and synthesis, retaining one NotebookLM conversation for context. An environment checker validates Python, dependencies, commands, MCP configuration, and NotebookLM authentication.
Tradeoffs
This is not a self-contained converter. NotebookLM authentication is mandatory, podcast transcription needs Get笔记 credentials and a separate web token, and WeChat extraction requires Playwright plus manual MCP configuration. Some archived paywall pages can stop for human verification. The paywall workflow impersonates crawlers and alters request headers, so users must assess publisher terms and authorization themselves. Documentation claims broader format handling than the supplied main program excerpt can verify.
How to use it well
It suits Claude Code users who repeatedly turn research material into NotebookLM notebooks, presentations, audio, or structured study notes. Start with public webpages, YouTube links, Markdown, PDF, or EPUB, then add browser and transcription integrations only when needed. Run the environment checks before real work and inspect extracted text before generation. It does not replace source-access rights, a general-purpose transcription service, or a tested document production pipeline independent of NotebookLM.
Technical notes+
main.py performs type detection, EPUB extraction, NotebookLM CLI subprocess calls, and a three-round, twelve-question analysis workflow that writes JSON under /tmp. scripts/get_podcast_transcript.py polls the Get笔记 OpenAPI, refreshes a JWT from ~/.claude/skills/getnote/tokens.json, and writes transcript text to a temporary file. requirements.txt declares FastMCP, Playwright, Beautiful Soup, lxml, and MarkItDown, while install.sh additionally installs Chromium and notebooklm-py. feishu-read-mcp/src/server.py annotates an argument with Optional but does not import that name, so the module is likely to fail during definition. feishu-read-mcp/test.py exercises imports, scraper initialization, parsing, and image handling as a standalone script.
Observed
- License
- MIT License
- Primary language
- Python, with Bash installation and fetching scripts
- Runtime
- Python 3.9 or newer
- Install surface
- Clone into the Claude skills directory, then run the included Bash installer
- Interfaces
- Claude Code skill, Python CLI, NotebookLM CLI integration, and FastMCP servers
- Core dependencies
- FastMCP, Playwright, Beautiful Soup, lxml, and MarkItDown
- External services
- NotebookLM is required; Get笔记 is optional for supported podcast and video transcription
- Testing structure
- A standalone Feishu MCP test script is present
Read from README.md, requirements.txt, main.py, check_env.py, feishu-read-mcp/test.py, scripts/get_podcast_transcript.py, feishu-read-mcp/src/server.py, feishu-read-mcp/src/parser.py, feishu-read-mcp/src/scraper.py, feishu-read-mcp/src/__init__.py, feishu-read-mcp/src/image_handler.py, LICENSE, SKILL.md, install.sh.
Tech Stack
Comments (0)
No comments yet
Editorially curated, with community endorsements as a secondary signal. Corrections welcome.