notebooklm
Skill1 starsChanged 6 months agoMITView on GitHub
Complete API for Google NotebookLM - full programmatic access including features not in the web UI. Create notebooks, add sources, generate all artifact types, download in multiple formats. Activates on explicit /notebooklm or intent like "create a podcast about X", "install notebooklm", "add notebooklm to cowork
- Deletes or force-pushes
- Installs packages
69 other projects use a version of this, 24 of them edited.
Adopted by
Most recently taken up by
- proyecto26/notebooklm-ai-plugintheir own version, first seen Nov 8, 2026
- LuisSambrano/antigravity-configtheir own version, first seen Nov 8, 2026
- delorenj/skillstheir own version, first seen Nov 7, 2026
- roomi-fields/notebooklm-skilltheir own version, first seen Nov 6, 2026
- TsinHzl/claude-code-configtheir own version, first seen Nov 1, 2026
- Mathews-Tom/armoryedited, first seen Nov 1, 2026
Used with
Projects with it also use django-verification, django-security and 10 more.
Other files people use alongside this one.
Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
mstuder-slalomSkilldjango-verification
Verification loop for Django projects: migrations, linting, tests with coverage, security scans, and deployment readiness checks before release or PR.
Also in 3 projects that use this
mstuder-slalomSkilldjango-security
Django security best practices, authentication, authorization, CSRF protection, SQL injection prevention, XSS prevention, and secure deployment configurations.
Also in 3 projects that use this
sirakinbSkilllast30days
Research a topic from the last 30 days on Reddit + X + Web, become an expert, and write copy-paste-ready prompts for the user's target tool.
Also in 4 projects that use this
younghaiSkillgoogle-calendar
Google Calendar 일정 조회, 검색, 등록. "일정", "스케줄", "캘린더" 등을 언급하면 자동 실행. Daily Note 작성, Weekly Review에 사용.
Also in 3 projects that use this
royaluniondesign-sysSkilldeployment-patterns
Deployment workflows, CI/CD pipeline patterns, Docker containerization, health checks, rollback strategies, and production readiness checklists for web applications.
Also in 3 projects that use this
LuisSambranoSkillfirecrawl
Deep web scraping, screenshots, PDF parsing, and website crawling using Firecrawl API
Also in 3 projects that use this
Other versions69 projects use a version of this.
Word for word · 0
None found.
Edited · 24
- Mathews-Tom/armory★ 34020% same
- ttmouse/skills★ 3411% same
- artguru-tech-team/claude-skills-public★ 923% same
- audibleblink/skills★ 436% same
- patriarchal-boothose896/notebooklm-py★ 323% same
- alfredang/skills★ 217% same
- leonardoaraujosantos/my_ai_skills★ 123% same
- VINK77aaaa/dsh-skills★ 117% same
- 3akat-system/nexus-starter★ 023% same
- alexhdsale/claude-skills★ 023% same
14 more not listed
Wrote their own version · 45
- teng-lin/notebooklm-py★ 20ksame name
- ArtemXTech/personal-os-skills★ 545same name
- sanjay3290/ai-skills★ 452same name
- roomi-fields/notebooklm-mcp★ 241same name
- cclank/notebooklm-skill★ 140same name
- chendongqi/OPB-Skills★ 125same name
- walidboulanouar/Ay-Skills★ 90same name
- proyecto26/notebooklm-ai-plugin★ 65same name
- artemnovitckii/notebooklm-coach★ 61same name
- rainday/smart-blog-skills★ 35same name
35 more not listed
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository16 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- speckit-analyzeSkill
- speckit-checklistSkill
- speckit-clarifySkill
- speckit-constitutionSkill
- speckit-git-commitSkill
- speckit-git-featureSkill
- speckit-git-initializeSkill
- speckit-git-remoteSkill
- speckit-git-validateSkill
- speckit-implementSkill
- speckit-planSkill
- speckit-specifySkill
- speckit-tasksSkill
- speckit-taskstoissuesSkill
The file
Show its rules and text
We split this skill into the rules it gives an agent
16 rules. 0 are common, 11 are distinctive, and 5 are very specific to this project.
0
Common
Most of the 66 edited versions we read keep these.
11
Distinctive
This author's own choices. Worth a look.
5
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 22
When this skill is triggered and notebooklm is not yet installed or authenticated, complete setup first.
- Distinctive: line 32
If Python is below 3.10 (e.g.
- Distinctive: line 47
Always use a virtual environment to avoid "externally-managed-environment" errors and PATH issues.
- Distinctive: line 72
Verify the CLI works:
- Distinctive: line 135
Wait ~10 seconds for the browser to open, then ask the user if they can see the browser and are signed in.
- Distinctive: line 153
If auth passes (SID cookie present), confirm to the user that NotebookLM is set up and ready.
- Distinctive: line 159
If auth fails (SID cookie missing), the user may not have fully signed in.
- Distinctive: line 173
When the user asks to "add this to Cowork", "use this in Cowork", or "make this work in Cowork":
- Distinctive: line 205
If either check fails, run the Step 0: Local Setup above before continuing.
- Distinctive: line 211
Create the project and set up a Python 3.12 virtual environment:
- Very specific: line 220
Create ~/notebooklm-mcp/server.py.
- Distinctive: line 489
All generate commands support:
- Very specific: line 513
Wait for sources: notebooklm source list --json until all status=READY
- Very specific: line 515
Check notebooklm artifact list for status
- Distinctive: line 525
Create one permanent notebook called something like "AI Brain"
- Very specific: line 526
After every working session, append a summary as a new source (the /wrapup skill automates this)
“Common” is measured: kept by at least 6 in 10 of the 66 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
<!-- notebooklm-py v0.3.4 -->NotebookLM Automation
Complete programmatic access to Google NotebookLM—including capabilities not exposed in the web UI. Create notebooks, add sources (URLs, YouTube, PDFs, audio, video, images), chat with content, generate all artifact types, and download results in multiple formats.
This skill has two setup paths:
- Local Claude Code — Install the
notebooklm-pyCLI on your Mac and authenticate. Fast, direct, no network overhead. Use this for terminal sessions. - Cowork — Wrap the CLI in an MCP server, tunnel it via Cloudflare, and connect Cowork to the public HTTPS endpoint. Use this when you want NotebookLM access inside Cowork (Anthropic's sandbox can't run your local CLI directly).
The original cookie-inlining approach for Cowork has been replaced with the MCP pattern below. It's more robust (no token expiry pain), secure (credentials never leave your Mac), and works for any local CLI you want to expose to Cowork.
Step 0: Local Setup (Run Automatically on First Use)
When this skill is triggered and notebooklm is not yet installed or authenticated, complete setup first.
Pre-flight: Check Python Version
notebooklm-py requires Python 3.10+. Check the available version before installing:
python3 --versionIf Python is below 3.10 (e.g. 3.9.x which is the macOS default), install a compatible version:
macOS (Homebrew):
brew install python@3.12Then use /opt/homebrew/bin/python3.12 (Apple Silicon) or /usr/local/bin/python3.12 (Intel) for the venv below.
Linux (apt):
sudo apt update && sudo apt install -y python3.12 python3.12-venvInstall the CLI
Always use a virtual environment to avoid "externally-managed-environment" errors and PATH issues.
Determine which Python to use — if the system python3 is 3.10+, use it directly. Otherwise use the one you just installed (e.g. python3.12):
# Set PYTHON to the correct binary (adjust if needed)
PYTHON=$(command -v python3.12 2>/dev/null || command -v python3.11 2>/dev/null || command -v python3.10 2>/dev/null || command -v python3)
# Verify it's 3.10+
$PYTHON -c "import sys; assert sys.version_info >= (3,10), f'Python {sys.version} is too old — need 3.10+'; print(f'Using Python {sys.version}')"
# Create venv and install
$PYTHON -m venv ~/.notebooklm-venv
source ~/.notebooklm-venv/bin/activate
pip install "notebooklm-py[browser]"
playwright install chromiumThen symlink so it's always on PATH:
mkdir -p ~/bin
ln -sf ~/.notebooklm-venv/bin/notebooklm ~/bin/notebooklm
export PATH="$HOME/bin:$PATH"Verify the CLI works:
notebooklm --helpAuthenticate
IMPORTANT: The built-in notebooklm login command requires interactive terminal input (pressing Enter after sign-in). Claude Code's bash tool does NOT support interactive input, so notebooklm login will fail — the browser opens and closes instantly. Instead, use this custom login script.
Tell the user:
I'm going to open a browser window — just sign into your Google account and navigate to notebooklm.google.com. Take your time, I'll wait for you to confirm before closing it.
Then write and run this login script:
cat > /tmp/nlm_login.py << 'PYEOF'
import json, os, time
from pathlib import Path
from playwright.sync_api import sync_playwright
STORAGE_PATH = Path.home() / ".notebooklm" / "storage_state.json"
PROFILE_PATH = Path.home() / ".notebooklm" / "browser_profile"
SIGNAL_FILE = Path("/tmp/nlm_save_signal")
SIGNAL_FILE.unlink(missing_ok=True)
STORAGE_PATH.parent.mkdir(parents=True, exist_ok=True)
print("Opening browser for Google login...")
print("Sign in to Google and navigate to notebooklm.google.com")
with sync_playwright() as p:
browser = p.chromium.launch_persistent_context(
user_data_dir=str(PROFILE_PATH),
headless=False,
args=["--disable-blink-features=AutomationControlled"],
)
page = browser.pages[0] if browser.pages else browser.new_page()
page.goto("https://notebooklm.google.com/")
print("Browser is open. Waiting for save signal...")
while not SIGNAL_FILE.exists():
time.sleep(1)
print("Save signal received! Capturing session...")
storage = browser.storage_state()
with open(STORAGE_PATH, "w") as f:
json.dump(storage, f)
cookie_names = [c["name"] for c in storage.get("cookies", [])]
print(f"Saved {len(cookie_names)} cookies: {cookie_names}")
browser.close()
SIGNAL_FILE.unlink(missing_ok=True)
print(f"Authentication saved to: {STORAGE_PATH}")
PYEOF
# Run the login script in the background
source ~/.notebooklm-venv/bin/activate
python3 /tmp/nlm_login.py > /tmp/nlm_login_output.txt 2>&1 &
echo "Login started (PID=$!). Browser should open in a few seconds..."Wait ~10 seconds for the browser to open, then ask the user if they can see the browser and are signed in.
Once the user confirms they are on the NotebookLM homepage, save the session:
touch /tmp/nlm_save_signal
sleep 8
cat /tmp/nlm_login_output.txtThen verify authentication:
export PATH="$HOME/bin:$PATH"
notebooklm auth check
notebooklm listIf auth passes (SID cookie present), confirm to the user that NotebookLM is set up and ready. Clean up the temp script:
rm -f /tmp/nlm_login.py /tmp/nlm_login_output.txt /tmp/nlm_save_signalIf auth fails (SID cookie missing), the user may not have fully signed in. Delete the browser profile and retry:
rm -rf ~/.notebooklm/browser_profile ~/.notebooklm/storage_state.jsonThen run the login script again from the top.
Adding NotebookLM to Cowork (MCP Server + Cloudflare Tunnel)
Credit: This approach was contributed by Daniel at skool.com/navaigate. It supersedes the older cookie-inlining method, which broke whenever Google rotated tokens.
When the user asks to "add this to Cowork", "use this in Cowork", or "make this work in Cowork":
Why the MCP Approach
Cowork runs on Anthropic's servers. It can't see your local files. It can't run your CLI tools. The old approach of inlining auth cookies into a skill file failed repeatedly because:
- Tokens expire and force a full regeneration cycle
- Google's auth flow can change without warning
- You have to re-upload the skill every time
The MCP pattern solves this cleanly: the CLI stays on your Mac (full filesystem access, full auth), you wrap it in an MCP server, Cloudflare Tunnel gives it a public HTTPS URL, and Cowork connects to that URL like any other MCP server. The agent doesn't know or care where the tool runs — it just works.
Architecture
Claude Code / Cowork
↓ MCP Protocol (SSE)
Cloudflare Tunnel (HTTPS)
↓
Your Mac — localhost:8484
↓
MCP Server (Python + FastMCP)
↓
NotebookLM CLI → Google NotebookLMPrerequisite Check
# Make sure the local CLI is installed and authenticated first
cat ~/.notebooklm/storage_state.json > /dev/null 2>&1 && notebooklm auth checkIf either check fails, run the Step 0: Local Setup above before continuing.
You'll also need a domain on Cloudflare (for the tunnel hostname). If the user doesn't have one yet, they'll need to set that up first at cloudflare.com.
Step 1: Build the MCP Server
Create the project and set up a Python 3.12 virtual environment:
mkdir -p ~/notebooklm-mcp && cd ~/notebooklm-mcp
/opt/homebrew/bin/python3.12 -m venv .venv
source .venv/bin/activate
pip install "mcp[cli]" uvicornCreate ~/notebooklm-mcp/server.py. Every CLI command becomes a decorated Python function — here's the pattern (extend with as many tools as you want exposed):
import sys, subprocess
from mcp.server.fastmcp import FastMCP
NOTEBOOKLM = "/Users/YOUR_USERNAME/bin/notebooklm" # adjust to your symlink path
mcp = FastMCP("notebooklm", host="0.0.0.0", port=8484)
def run_cli(*args, timeout=120):
result = subprocess.run(
[NOTEBOOKLM, *args],
capture_output=True, text=True, timeout=timeout
)
return result.stdout.strip() or "(no output)"
@mcp.tool()
def notebooklm_list() -> str:
"""List all NotebookLM notebooks."""
return run_cli("list")
@mcp.tool()
def notebooklm_use(notebook_id: str) -> str:
"""Set the active notebook context."""
return run_cli("use", notebook_id)
@mcp.tool()
def notebooklm_ask(question: str) -> str:
"""Ask the current notebook a question (RAG query)."""
return run_cli("ask", question, timeout=120)
@mcp.tool()
def notebooklm_source_add(url_or_path: str) -> str:
"""Add a URL, YouTube link, or local file as a source to the current notebook."""
return run_cli("source", "add", url_or_path, timeout=180)
@mcp.tool()
def notebooklm_generate_audio(instructions: str = "") -> str:
"""Generate a podcast-style audio overview."""
args = ["generate", "audio"]
if instructions:
args.append(instructions)
return run_cli(*args, timeout=600)
# Follow this pattern for every command you want to expose —
# create, source list, generate video/report/quiz/flashcards, download, etc.
# See the Quick Reference table further down for the full command surface.
if __name__ == "__main__":
if "--sse" in sys.argv:
import uvicorn
app = mcp.sse_app()
uvicorn.run(app, host="0.0.0.0", port=8484)
else:
mcp.run(transport="stdio")The entry point supports two transport modes:
- SSE mode (
--sse) — for Cowork via the tunnel. Runs an HTTP server on port 8484. - Stdio mode (default) — for local Claude Code. Direct pipe, no network overhead.
Smoke test it:
source ~/notebooklm-mcp/.venv/bin/activate
python -c "from server import run_cli; print(run_cli('auth', 'check'))"Step 2: Set Up the Cloudflare Tunnel
Cowork runs on Anthropic's servers. localhost means nothing to it. You need a public HTTPS URL pointing at your Mac. Cloudflare Tunnel does this without opening any ports on your router.
brew install cloudflared
cloudflared tunnel login
cloudflared tunnel create notebooklm-mcp
cloudflared tunnel route dns notebooklm-mcp mcp-notebooklm.yourdomain.comCreate ~/.cloudflared/config-notebooklm-mcp.yml:
tunnel: YOUR_TUNNEL_ID
credentials-file: ~/.cloudflared/YOUR_TUNNEL_ID.json
ingress:
- hostname: mcp-notebooklm.yourdomain.com
service: http://localhost:8484
- service: http_status:404Test it end-to-end:
# Terminal 1: start the server
cd ~/notebooklm-mcp && source .venv/bin/activate
python server.py --sse
# Terminal 2: start the tunnel
cloudflared tunnel --config ~/.cloudflared/config-notebooklm-mcp.yml run notebooklm-mcp
# Terminal 3: verify
curl -s https://mcp-notebooklm.yourdomain.com/sse | head -3You should see event: endpoint with a session ID. That means it's working.
Step 3: Auto-Start on Boot (macOS Launch Agents)
You don't want to manually start two processes every time you turn on your Mac. Create two Launch Agents:
MCP Server — ~/Library/LaunchAgents/dev.navaigate.notebooklm-mcp.plist
Points to .venv/bin/python server.py --sse. Set RunAtLoad and KeepAlive to true.
Tunnel — ~/Library/LaunchAgents/dev.navaigate.notebooklm-tunnel.plist
Points to cloudflared tunnel run with the config file. Same flags.
Load them once — they persist across reboots, and macOS restarts either process automatically if it crashes:
launchctl load ~/Library/LaunchAgents/dev.navaigate.notebooklm-mcp.plist
launchctl load ~/Library/LaunchAgents/dev.navaigate.notebooklm-tunnel.plistStep 4: Connect to Cowork
- Open Cowork and click the + icon
- Go to Connectors → Add connection → Add custom connector
- Enter:
https://mcp-notebooklm.yourdomain.com/sse - Click Add
The NotebookLM tools appear immediately. Prompt naturally:
"List my NotebookLM notebooks" "Switch to my AI Brain notebook and tell me the last 3 entries" "Ask my AI Brain: what were the key decisions from last week?" "Generate a podcast from my current notebook"
Step 5: Local Claude Code Setup (Stdio Transport)
For terminal sessions you don't need the tunnel — add the MCP server to ~/.claude/settings.json using stdio transport:
{
"mcpServers": {
"notebooklm": {
"command": "/Users/YOUR_USERNAME/notebooklm-mcp/.venv/bin/python",
"args": ["/Users/YOUR_USERNAME/notebooklm-mcp/server.py"]
}
}
}Same server, different transport. The tools work identically — local sessions just skip the network hop.
MCP Troubleshooting
| Symptom | Fix |
| Auth expired | Run notebooklm login (or the custom login script in Step 0) locally. The MCP server picks up new tokens automatically — no restart needed. |
| Server not responding | lsof -i :8484 to check the port, tail ~/notebooklm-mcp/mcp-server.log for errors, then launchctl kickstart -k gui/$(id -u)/dev.navaigate.notebooklm-mcp. |
| Tunnel down | cloudflared tunnel info notebooklm-mcp. KeepAlive usually restarts it automatically. |
| Cowork says "can't reach server" | Verify with curl -s https://mcp-notebooklm.yourdomain.com/sse | head -3. If that works, remove and re-add the connector in Cowork. |
Why This Pattern Generalizes
Local MCP server + Cloudflare Tunnel isn't just for NotebookLM. Any CLI tool, any local service, any API that needs local credentials can be wrapped as an MCP server and made available to Cowork the same way. Once you've done it once, you can do it for anything.
Resources
- NotebookLM CLI Skill for Claude Code: github.com/skyremote/claude-code-notebooklm-skills
- notebooklm-py (CLI tool): github.com/jgravelle/notebooklm-py
- MCP Python SDK: github.com/modelcontextprotocol/python-sdk
- Cloudflare Tunnel docs: developers.cloudflare.com/cloudflare-one/connections/connect-networks
When This Skill Activates
Explicit: User says "/notebooklm", "use notebooklm", "install notebooklm", or mentions the tool by name.
Intent detection: Recognize requests like:
- "Create a podcast about [topic]"
- "Summarize these URLs/documents"
- "Generate a quiz from my research"
- "Turn this into an audio overview"
- "Create flashcards for studying"
- "Generate a video explainer"
- "Make an infographic"
- "Create a mind map of the concepts"
- "Download the quiz as markdown"
- "Add these sources to NotebookLM"
- "Add this to Cowork" / "Make this work in Cowork" → use the MCP pattern above
Autonomy Rules
Run automatically (no confirmation):
notebooklm status- check contextnotebooklm auth check- diagnose auth issuesnotebooklm list- list notebooksnotebooklm source list- list sourcesnotebooklm artifact list- list artifactsnotebooklm language list- list supported languagesnotebooklm language get- get current languagenotebooklm language set- set language (global setting)notebooklm artifact wait- wait for artifact completionnotebooklm source wait- wait for source processingnotebooklm research status- check research statusnotebooklm research wait- wait for researchnotebooklm use <id>- set contextnotebooklm create- create notebooknotebooklm ask "..."- chat queries (without--save-as-note)notebooklm history- display conversation history (read-only)notebooklm source add- add sources
Ask before running:
notebooklm delete- destructivenotebooklm generate *- long-running, may failnotebooklm download *- writes to filesystemnotebooklm ask "..." --save-as-note- writes a notenotebooklm history --save- writes a note
Quick Reference
| Task | Command |
| List notebooks | notebooklm list |
| Create notebook | notebooklm create "Title" |
| Set context | notebooklm use <notebook_id> |
| Show context | notebooklm status |
| Add URL source | notebooklm source add "https://..." |
| Add file | notebooklm source add ./file.pdf |
| Add YouTube | notebooklm source add "https://youtube.com/..." |
| List sources | notebooklm source list |
| Wait for source processing | notebooklm source wait <source_id> |
| Web research (fast) | notebooklm source add-research "query" |
| Web research (deep) | notebooklm source add-research "query" --mode deep --no-wait |
| Check research status | notebooklm research status |
| Wait for research | notebooklm research wait --import-all |
| Chat | notebooklm ask "question" |
| Chat (specific sources) | notebooklm ask "question" -s src_id1 -s src_id2 |
| Chat (with references) | notebooklm ask "question" --json |
| Chat (save answer as note) | notebooklm ask "question" --save-as-note |
| Show conversation history | notebooklm history |
| Save all history as note | notebooklm history --save |
| Get source fulltext | notebooklm source fulltext <source_id> |
| Generate podcast | notebooklm generate audio "instructions" |
| Generate video | notebooklm generate video "instructions" |
| Generate report | notebooklm generate report --format briefing-doc |
| Generate quiz | notebooklm generate quiz |
| Generate flashcards | notebooklm generate flashcards |
| Generate infographic | notebooklm generate infographic |
| Generate mind map | notebooklm generate mind-map |
| Generate slide deck | notebooklm generate slide-deck |
| Revise a slide | notebooklm generate revise-slide "prompt" --artifact <id> --slide 0 |
| Check artifact status | notebooklm artifact list |
| Wait for completion | notebooklm artifact wait <artifact_id> |
| Download audio | notebooklm download audio ./output.mp3 |
| Download video | notebooklm download video ./output.mp4 |
| Download slide deck (PDF) | notebooklm download slide-deck ./slides.pdf |
| Download slide deck (PPTX) | notebooklm download slide-deck ./slides.pptx --format pptx |
| Download report | notebooklm download report ./report.md |
| Download mind map | notebooklm download mind-map ./map.json |
| Download data table | notebooklm download data-table ./data.csv |
| Download quiz | notebooklm download quiz quiz.json |
| Download flashcards | notebooklm download flashcards cards.json |
| List languages | notebooklm language list |
| Set language | notebooklm language set zh_Hans |
Generation Types
All generate commands support:
-s, --sourceto use specific source(s) instead of all sources--languageto set output language (defaults to 'en')--jsonfor machine-readable output--retry Nto automatically retry on rate limits
| Type | Command | Options | Download |
| Podcast | generate audio | --format [deep-dive|brief|critique|debate], --length [short|default|long] | .mp3 |
| Video | generate video | --format [explainer|brief], --style [auto|classic|whiteboard|kawaii|anime|watercolor|retro-print|heritage|paper-craft] | .mp4 |
| Slide Deck | generate slide-deck | --format [detailed|presenter], --length [default|short] | .pdf / .pptx |
| Slide Revision | generate revise-slide "prompt" --artifact <id> --slide N | --wait, --notebook | (re-downloads parent deck) |
| Infographic | generate infographic | --orientation [landscape|portrait|square], --detail [concise|standard|detailed] | .png |
| Report | generate report | --format [briefing-doc|study-guide|blog-post|custom], --append "extra instructions" | .md |
| Mind Map | generate mind-map | (sync, instant) | .json |
| Data Table | generate data-table | description required | .csv |
| Quiz | generate quiz | --difficulty [easy|medium|hard], --quantity [fewer|standard|more] | .json/.md/.html |
| Flashcards | generate flashcards | --difficulty [easy|medium|hard], --quantity [fewer|standard|more] | .json/.md/.html |
Common Workflows
Research to Podcast
notebooklm create "Research: [topic]"notebooklm source addfor each URL/document- Wait for sources:
notebooklm source list --jsonuntil all status=READY notebooklm generate audio "Focus on [specific angle]"- Check
notebooklm artifact listfor status notebooklm download audio ./podcast.mp3when complete
Document Analysis
notebooklm create "Analysis: [project]"notebooklm source add ./doc.pdf(or URLs)notebooklm ask "Summarize the key points"- Continue chatting as needed
Second-Brain Pattern (the "AI Brain" workflow)
- Create one permanent notebook called something like "AI Brain"
- After every working session, append a summary as a new source (the
/wrapupskill automates this) - In Cowork, add a project instruction: "Whenever answering questions about strategy, business context, or project history, always consult my NotebookLM AI Brain notebook first."
- Claude now checks your second brain before answering — perfect recall of everything you've ever worked on
Output Formats (--json)
// notebooklm list --json
{"notebooks": [{"id": "...", "title": "...", "created_at": "..."}]}
// notebooklm source list --json
{"sources": [{"id": "...", "title": "...", "status": "ready|processing|error"}]}
// notebooklm artifact list --json
{"artifacts": [{"id": "...", "title": "...", "type": "Audio Overview", "status": "in_progress|pending|completed|unknown"}]}Error Handling
| Error | Cause | Action |
| Auth/cookie error | Session expired | Re-run the login script from Step 0 |
| "No notebook context" | Context not set | Run notebooklm use <id> |
| Rate limiting | Google throttle | Wait 5-10 min, retry |
| Download fails | Generation incomplete | Check artifact list for status |
| Cowork tool calls fail | MCP server or tunnel down | See MCP Troubleshooting section above |
Known Limitations
- Audio, video, quiz, flashcard, infographic, and slide deck generation may fail due to Google rate limits
- Generation times: audio 10-20 min, video 15-45 min, quiz/flashcards 5-15 min
- This is an unofficial API — Google can change things without warning
- The MCP server runs on your Mac, so Cowork access requires your Mac to be awake and online. For 24/7 access, deploy the server to a VPS and point the tunnel there instead.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

