skillZs
★ LIVE SKILL TAGS ★
>>> LIVE SKILLS INDEX <<<
* OPEN SOURCE *
NO LOGIN, NO TRACKING
※ REAL INSTALL DATA ※
← back to all skills
proyecto26/notebooklm-ai-plugin321 installs

notebooklm

Interact with Google NotebookLM (notebook.google.com) notebooks — chat with the AI, generate artifacts (slides, audio, video, mind maps, quizzes, flashcards, infographics, reports, data tables), manage sources (add URLs, YouTube, files, text), run research (fast/deep web research), and manage notes. Use whenever the user mentions NotebookLM, a notebook.google.com URL, audio overviews, or wants to query, create content from, or manage their notebook sources — even if they don't say "NotebookLM" explicitly.

How do I install this agent skill?

npx skills add https://github.com/proyecto26/notebooklm-ai-plugin --skill notebooklm
view source ↗

Is this agent skill safe to install?

  • Gen Agent Trust Hubpass

    This skill allows an AI agent to manage and interact with Google NotebookLM notebooks. It supports chatting with the notebook AI, conducting web research, and generating various artifacts like slide decks and audio overviews. To authenticate, it launches a local browser instance to capture session cookies. The primary security consideration is the ingestion of content from external URLs and files, which represents an attack surface for indirect prompt injection.

  • Socketwarn

    1 alert: gptAnomaly

  • Snykwarn

    Risk: MEDIUM · 1 issue

What does this agent skill do?

NotebookLM AI Plugin

Drive Google NotebookLM (https://notebook.google.com) from the terminal. Both notebook.google.com and notebooklm.google.com notebook URLs are accepted.

Supports:

  • Chat with Notebook AI (source-grounded Q&A with citations)
  • Slide Deck generation (PPTX/PDF)
  • Audio Overview (M4A — deep dive, brief, critique, debate formats)
  • Video Overview (MP4 — auto, classic, whiteboard, kawaii, anime, watercolor, retro_print, heritage, paper_craft styles)
  • Mind Map (JSON tree)
  • Flashcards / Quiz (interactive HTML)
  • Infographic (PNG — landscape, portrait, square; 11 visual styles)
  • Report (Markdown — briefing doc, study guide, blog post, custom prompt)
  • Data Table (CSV)
  • Source management (add URLs, YouTube, files, pasted text; list, delete)
  • Fast/Deep web research with auto-import
  • Notes management (create, update, delete, list)
  • Notebook library management (add, list, search, activate) + listing your account's notebooks

All commands run from this skill directory:

cd <skill-dir>   # the folder containing SKILL.md
npx -y bun scripts/main.ts <command> [options]

Quick start

# 1. Authenticate (opens Chrome for Google login; cookies are cached)
npx -y bun scripts/main.ts login

# 2. Check the session and see your notebooks (ids + source counts)
npx -y bun scripts/main.ts notebooks remote

# 3. Save one as the default
npx -y bun scripts/main.ts notebooks add https://notebook.google.com/notebook/<id> --name "My Research"

# 4. Use it
npx -y bun scripts/main.ts chat --question "What are the key findings?"
npx -y bun scripts/main.ts generate slide_deck --output slides.pdf
npx -y bun scripts/main.ts generate audio --format deep_dive --length long
npx -y bun scripts/main.ts generate quiz --difficulty medium --quantity more --output quiz.html
npx -y bun scripts/main.ts generate data_table --instructions "Compare the frameworks" --output table.csv

Authentication

login launches Chrome (with a dedicated profile under the data dir) and waits for you to sign in at notebook.google.com, then captures the Google session cookies. Cookies expire after a few weeks; any command that hits a login redirect tells you to run login --force.

npx -y bun scripts/main.ts login            # reuse cached cookies if still valid
npx -y bun scripts/main.ts login --force    # always re-open Chrome
npx -y bun scripts/main.ts notebooks remote # fast auth check: lists notebooks from your account

Every RPC call fetches a fresh CSRF token (SNlM0e) and session id (FdrFJe) from the app page, so nothing else needs refreshing. If the origin changes again, set NOTEBOOKLM_BASE_URL (or edit scripts/constants.ts).

Commands

Notebook management

npx -y bun scripts/main.ts notebooks remote                 # notebooks on your Google account (live)
npx -y bun scripts/main.ts notebooks list                   # local library
npx -y bun scripts/main.ts notebooks add <url> [--name <name>] [--description <desc>] [--topics <t1,t2>]
npx -y bun scripts/main.ts notebooks remove <id>
npx -y bun scripts/main.ts notebooks activate <id>          # default for every other command
npx -y bun scripts/main.ts notebooks search <query>

Chat with Notebook AI

npx -y bun scripts/main.ts chat --question "What are the key findings?" --notebook <url|id>
npx -y bun scripts/main.ts chat --question "Tell me more about that" --conversation-id <id>
npx -y bun scripts/main.ts chat --question "Summarize the methodology" --json   # answer + citations

Chat uses all ready sources in the notebook. Citations reference source ids (see sources list).

Source management

npx -y bun scripts/main.ts sources list --notebook <url|id>          # id, title, type, status, url
npx -y bun scripts/main.ts sources add-url https://example.com/article
npx -y bun scripts/main.ts sources add-youtube https://youtube.com/watch?v=xxx
npx -y bun scripts/main.ts sources add-text --title "My Notes" --content "Important findings..."
npx -y bun scripts/main.ts sources add-file ./paper.pdf
npx -y bun scripts/main.ts sources delete <sourceId>

New sources show status: processing for a little while; generation and chat only use sources whose status is ready. Supported files: PDF, TXT, MD, DOCX, PPTX, CSV, EPUB, images (PNG, JPG, WEBP), audio/video.

Research

npx -y bun scripts/main.ts research fast --query "latest AI agent frameworks"   # ~1–3 min
npx -y bun scripts/main.ts research deep --query "state of LLM reasoning"       # ~5–10 min, produces a report
npx -y bun scripts/main.ts research fast --query "topic" --import                # import found sources
npx -y bun scripts/main.ts research status

Notes

npx -y bun scripts/main.ts notes list
npx -y bun scripts/main.ts notes create --title "Key Takeaways" --content "1. Finding one..."
npx -y bun scripts/main.ts notes update <noteId> --title "Updated" --content "New content"
npx -y bun scripts/main.ts notes delete <noteId>

Artifacts

npx -y bun scripts/main.ts artifacts list                 # everything in the Studio panel, with status
npx -y bun scripts/main.ts generate <type> [options]

Types: slide_deck, audio, video, mind_map, flashcards, quiz, infographic, report, data_table

Generation is asynchronous on Google's side: the command creates the artifact, polls the Studio list until it is completed and its download URL is available (audio/video can take 10–20 minutes), then downloads it. If auto-download fails, the URL is printed so it can be opened in a browser.

Options

Global options

OptionDescription
--notebook <url|id>Notebook URL or library id (defaults to the active notebook)
--output <path>Output file path (auto-named under the data dir if omitted)
--instructions <text>Custom instructions / focus for generation
--language <code>Output language (default en)
--jsonMachine-readable output

Per-type options

TypeOptions
audio--format deep_dive|brief|critique|debate, --length short|default|long
video--style auto|classic|whiteboard|kawaii|anime|watercolor|retro_print|heritage|paper_craft, --format explainer|brief|cinematic|short
slide_deck--format detailed|presenter|pdf|pptx, --length default|short (PPTX by default; --format pdf or an .pdf output path selects the PDF)
quiz, flashcards--difficulty easy|medium|hard, --quantity fewer|standard|more
infographic--orientation landscape|portrait|square, --detail concise|standard|detailed, --style auto|sketch_note|professional|bento_grid|editorial|instructional|bricks|clay|anime|kawaii|scientific
report--format briefing|study_guide|blog_post|custom, --prompt "<full prompt>" (custom)
mind_map--instructions only
data_table--instructions (what to tabulate)

Output formats

TypeOutput
slide_deckPPTX (or PDF)
audioM4A
videoMP4
mind_mapJSON tree { name, children: [...] }
flashcards, quizself-contained interactive HTML
infographicPNG
reportMarkdown
data_tableCSV (headers + rows)

Environment variables

VariableDescription
NOTEBOOKLM_BASE_URLApp origin override (default https://notebook.google.com)
NOTEBOOKLM_DATA_DIRData directory (cookies, library, Chrome profile, outputs)
NOTEBOOKLM_COOKIE_PATHCookie file path
NOTEBOOKLM_CHROME_PATH / NOTEBOOKLM_CHROME_PROFILE_DIRChrome executable / profile
NOTEBOOKLM_OUTPUT_DIRDefault output directory
NOTEBOOKLM_BLBuild label sent on chat requests (not validated by the server)

Rate limits (free tier)

ResourceLimit
Audio/video overviews3 per day
Reports/flashcards/quizzes10 per day
Daily chats50
Notebooks100, 50 sources each

Quota rejections surface as an RPCError of kind rate_limit.

Troubleshooting

SymptomMeaning / fix
redirected (302) to …/loginCookies expired → login --force
redirected … The app origin may have changedGoogle moved the app again → set NOTEBOOKLM_BASE_URL
RPC xxxx not present in response … method ID may have changedGoogle rotated an RPC id → update scripts/rpc-types.ts (compare with notebooklm-py)
INVALID_ARGUMENT / FAILED_PRECONDITIONPayload shape rejected — run npx -y bun test and diff scripts/*-manager.ts against the reference projects
did not return an artifact IDNotebook has no ready sources, or the artifact type isn't enabled for the account

Tests

cd test
npx -y bun install                          # dev deps only (typescript, types)
npx -y bun x tsc --noEmit                   # type-check
npx -y bun test unit                        # offline: payload shapes + decoders against recorded responses
npx -y bun test live.test.ts                # live: auth + read endpoints (skips without a session)
NOTEBOOKLM_LIVE_WRITE=1 npx -y bun test live.test.ts             # + add/delete a text source and a note
NOTEBOOKLM_LIVE_GENERATE=quiz npx -y bun test live.test.ts       # + generate one artifact end-to-end

Skill-level evals (prompts + assertions for the skill-creator harness) live in test/evals.json.

Add the canonical catalog link to the repository README so users can inspect current installs and available audits. The publishing guide covers the complete discovery path.

<a href="https://skillzs.dev/skills/proyecto26/notebooklm-ai-plugin/notebooklm">View notebooklm on skillZs</a>