Automate NotebookLM content creation

Control Google NotebookLM via browser automation

Users who want to programmatically control NotebookLM (read, add sources, generate outputs) without manual browser interaction.

• · NotebookLM action (read notebook, add source, generate Studio output, create new)
• · Notebook name (for read), source details (for add), Studio output type (for generate)

• · Notebook content (if reading)
• · Studio output (audio overview, infographics, slide decks, study guides, briefing docs, mind maps, timelines, FAQs)
• · New notebook creation confirmation

• · Browser automation (Playwright or similar)
• · NotebookLM web interface

• · Browser automation environment must be available (fails gracefully if unavailable)
• · NotebookLM account must be authenticated (browser session must persist)
• · Supported source types: URLs, text, files, YouTube links, synthesized content

• · Browser automation can timeout on slow networks (no retry built in)
• · NotebookLM UI changes can break selectors (fragile to layout updates)
• · Studio output generation is slow (5-10 sec per output, no progress indication)

• · Source spec in PR #657 (megaprompts/03-notebooklm-megaprompt.md)
• · Browser-automation shape differs from research-pack convention (action-routing intake)
• · Supports all NotebookLM Studio outputs (7 types documented)
• · Graceful failure when browser automation unavailable (CLI-only portability)