Skip to content
SF CTA Study Hub

NotebookLM Study Guide

Overview

NotebookLM is a Google AI tool that ingests your source documents and generates study materials grounded in those documents — quizzes, flashcards, audio overviews, mind maps, and Q&A responses. Unlike general-purpose AI tools, NotebookLM answers only from what you uploaded, making it accurate for domain-specific exam prep.

For CTA study the workflow is: upload one domain’s docs as sources → generate study materials → use Q&A to test understanding. Each CTA domain gets its own notebook so generated materials stay focused and relevant.

The batch upload script at scripts/notebooklm-upload.sh automates the setup: it creates 8 per-domain notebooks and loads the matching docs/ files as sources.

Prerequisites

1. Create a Google account (if needed)

NotebookLM requires a Google account. Visit notebooklm.google.com to verify you have access.

2. Authenticate the CLI

Terminal window
notebooklm login

This opens a browser window for Google OAuth. After completing login, credentials are stored at ~/.notebooklm/storage_state.json. This is a one-time step — subsequent commands use the stored session.

Verify auth worked:

Terminal window
notebooklm list

If the command returns without error (even with an empty list), authentication succeeded.

Initial Setup — Batch Upload

Run the upload script from the repo root:

Terminal window
bash scripts/notebooklm-upload.sh

This script creates 8 notebooks and loads the matching docs/ files into each:

NotebookSource Files
CTA: 01 System Architecturedocs/01-system-architecture/index.md
CTA: 02 Securitydocs/02-security/index.md
CTA: 03 Data Architecturedocs/03-data/index.md
CTA: 04 Solution Architecturedocs/04-solution-architecture/index.md
CTA: 05 Integrationdocs/05-integration/index.md
CTA: 06 Dev Lifecycledocs/06-dev-lifecycle/index.md
CTA: 07 Communicationdocs/07-communication/index.md
CTA: Exam Overviewdocs/exam-overview/*.md (4 files)

After the script completes, verify with:

Terminal window
notebooklm list

You should see all 8 notebooks listed. If the script printed any WARNING: file not found lines, those source files do not yet exist in docs/ and can be added manually once created.

Study Workflow

Select a domain notebook

Terminal window
notebooklm use "CTA: 05 Integration"

Generate study materials

Terminal window
notebooklm generate quiz         # multiple-choice quiz based on sources
notebooklm generate flashcards   # Q&A flashcard set
notebooklm generate audio        # ~10 min audio overview (podcast-style)
notebooklm generate mind-map     # visual concept map
notebooklm generate report       # detailed study summary

Generated materials appear in the NotebookLM web UI at notebooklm.google.com. They are not saved to local files.

Ask targeted questions

Terminal window
notebooklm ask "What are the key differences between Salesforce Shield and Classic Encryption?"
notebooklm ask "When should I use Platform Events vs Change Data Capture?"
notebooklm ask "Walk me through the CTA design defense scoring rubric"

Answers are grounded in the uploaded source documents — not general internet knowledge. This makes the Q&A reliable for verifying your understanding of specific concepts in the study materials.

Adding New Sources

When docs/ content is updated, add the changed file to the relevant notebook:

Terminal window
notebooklm use "CTA: 05 Integration"
notebooklm source add docs/05-integration/index.md

Or re-run the full upload script to refresh all notebooks:

Terminal window
bash scripts/notebooklm-upload.sh

Duplicate notebooks

Re-running the upload script calls notebooklm create again for each notebook. NotebookLM may create duplicates rather than overwriting. Run notebooklm list before re-running to check whether notebooks already exist, and delete duplicates from the NotebookLM web UI if needed.

Study Strategy Recommendations

GoalApproach
Cover a domain before practicegenerate audio — listen on the go
Active recall practicegenerate flashcards then self-test
Mock exam prepgenerate quiz per domain
Clarify a specific conceptask "explain X in the context of CTA"
Build mental modelgenerate mind-map to see concept relationships
Full domain summarygenerate report before a study session