Multi-source -> NotebookLM Processor

What it does

Automatically ingests content from many sources (WeChat articles, web pages, YouTube/B站, podcasts, PDFs, EPUB, DOCX, images with OCR, and local files), uploads them to NotebookLM, and optionally generates downstream artefacts such as podcasts (audio), slide decks (PPT), mind-maps, quizzes and written reports. It includes logic for paywall bypass strategies and integration with transcription services for podcasts/videos.

When to use it

Use this skill when you need to quickly convert heterogeneous content into NotebookLM sources or produce consumable formats (audio, slides, mind-maps, quizzes) from an article, video, podcast or document. Triggers include: a WeChat article URL, YouTube/B站 links, podcast links (xiaoyuzhou/ximalaya), local PDFs/EPUBs/DOCX, or natural-language requests like "make a podcast from this article".

What's included

• Scripts: none bundled, but the skill references helper scripts and MCP integrations for WeChat and GetNote transcription APIs.
• References: README/frontmatter points to the repo homepage and required third-party tools (NotebookLM CLI, GetNote API).
• Instructions: detailed workflow for source identification, content extraction, optional deep-analysis mode that generates recursive questions, and mapping of user intents to NotebookLM generate commands.

Compatible agents

Best suited for Claude Code / Claude Code-like agents with shell access and external CLI tooling (NotebookLM). The skill assumes access to local helper scripts and network APIs for transcription services.