notebooklm mcp
Automate Google NotebookLM to ingest sources, research topics, and generate podcasts, slides, and grounded reports.
Free
Works with the AI Agents you already use
notebooklm mcp
Example session with this skill installed
Create a notebook called "Project Aurora", add the PDF at ./specs/aurora_v2.pdf, and generate a long-form deep dive audio overview of it.
- Read your context and instructions
- Compiled the notebooklm mcp
I have initialized the "Project Aurora" notebook and uploaded your specification document. I am now triggering the audio generation process.
Notebook ID: nb_7829ac41
Source Added: aurora_v2.pdf (ID: src_9921)
Task Started: Audio Overview (Deep Dive, Long)
NotebookLM Task Summary
| Component | Status | Details |
|---|---|---|
| Notebook Name | Created | Project Aurora |
| Source Processing | Completed | 1 PDF successfully indexed |
| Generation Task | In Progress | Audio Overview (Deep Dive) |
| Estimated Time | ~15 Minutes | Task ID: task_audio_5521 |
The audio generation has been queued. Because this is a long-form "Deep Dive" format, it typically takes between 10 to 20 minutes for Google's servers to synthesize the conversation. I will monitor the status periodically.
Next steps
- Run
get_artifact_statusin 10 minutes to check if the audio is ready. - Use
download_artifactonce completed to save the file locally. - Ask specific questions about the
aurora_v2.pdfcontent to see citations while you wait.
Connects securely to your tools. The creator never sees your data.
About this skill
Managing document-heavy research manually is a bottleneck for analysis. This skill connects your agent to Google NotebookLM via the nblm-mcp server, allowing you to automate the ingestion of web pages, YouTube videos, and PDFs into structured notebooks for grounded analysis.
What it does
- Source ingestion imports web URLs, YouTube links, local PDFs, and raw text into managed notebooks.
- Grounded querying executes questions against your documents with precise citation mapping to source materials.
- Artifact generation automates the creation of deep-dive audio overviews, slide decks, study guides, and infographics.
- Automated research performs deep web or Google Drive surveys to populate notebooks with relevant citations.
- Content extraction generates structured data tables, mind maps, quizzes, and flashcards from unstructured sources.
How it works
- Connection check verifies the
nblm-mcpserver is active and the user is authenticated via thenotebooklm logincommand. - Notebook management creates or selects notebooks and adds sources using
add_url_sourceoradd_file_source. - Synthesis & generation triggers specific artifact tools like
generate_audio_overvieworgenerate_reportbased on the objective. - Status monitoring polls
get_artifact_statusfor long-running generation tasks before downloading the final file.
Frameworks & tools
Built for use with Claude Code or Claude Desktop. It requires the nblm-mcp server, notebooklm-py library, and a Chromium installation for authentication.
Why this beats prompting it yourself
Standard LLMs suffer from limited context windows and hallucinations when handling large document sets. This skill offloads the heavy lifting to Google's specialized RAG engine, providing verified citations and specialized studio outputs like audio and PPTX that a standard chat interface cannot produce.
Use cases
- Convert a technical documentation folder into a "deep dive" podcast for team onboarding.
- Analyze multiple PDF research papers to generate a structured data table of findings.
- Transform a YouTube lecture series into a set of interactive flashcards and a slide deck.
- Conduct deep research on a market trend and automatically import all cited web sources.
Known limitations
Requires a manual browser-based login for Google account authentication. Audio and video generation tasks can take up to 45 minutes and are subject to NotebookLM account quotas.
Requires manual browser login via terminal. Generation tasks can take 15-45 minutes. Subject to NotebookLM account rate limits. Audio/video tools may fail on headless environments without setup.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Free forever
No account required to browse
Trust & safety
Security scan
Security scan passed 16 days ago
- Free with an Agensi account
Requires