NotebookLM Audio Generator Skill

This skill automates the process of using Google's NotebookLM to generate a deep-dive audio overview from various sources and download it to a specified directory.

Prerequisites

This skill relies on the notebooklm-mcp-cli package and the epub2txt utility for .epub support. Before proceeding, ensure the user has them installed and authenticated.

Installation & Setup

If the user has not set up the CLI or the conversion tool, instruct them to do so first:

1. 1. Install the NotebookLM CLI:

   uv tool install notebooklm-mcp-cli
   # OR
   pip install notebooklm-mcp-cli
   

1. 2. Authenticate:

   nlm login
   

1. 3. Install epub2txt (for .epub support):

Workflow

When activated, follow these steps strictly in order. Do not skip steps.

Step 1: Gather Information

Ask the user for the following information one by one to prepare for the generation:

1. 1. Ask for a desired name for the new notebook. If not provided, use the default name "Audio Overview Notebook".
2. Source Selection: Ask the user to select their sources from the following options:

1. 3. Preferred Language: Ask the user for the preferred output language (BCP-47 code).

1. 4. Download Destination: Use a GUI directory picker to select where the audio file should be saved.

Crucial: Ask the user interactively to confirm they have provided/selected all the sources they wish to include before proceeding.

Step 2: Create a New Notebook

Use the notebook_create tool to create a new notebook with the provided name. Keep track of the notebook_id.

Step 3: Upload Sources

• - Loop through all gathered sources:

• - Always set wait=true to ensure sources are processed.

Step 4: Generate Audio Overview

• - Use the studio_create tool to start the audio generation.

Step 5: Monitor Generation Status

• - Inform the user: Explicitly tell the user that generating a long, deep-dive audio overview can take 5 to 15 minutes.
• Use the studio_status tool with action: status in a polling loop (using run_shell_command with sleep 300 between checks) until the status becomes completed.
• Note the audio_url and artifact_id when finished. Crucial: If there are multiple audio artifacts returned in the status, always identify and note the latest one (the one with the most recent created_at timestamp).

Step 6: Download the Audio

• - Using the artifact_id of the latest audio, try using the download_artifact tool first to save the audio to the destination path with an .mp3 extension.

Step 7: Final Verification

• - Verify the downloaded file using file <output_path>.
• Inform the user that the process is complete.