education

How to Use NotebookLM: Step-by-Step Guide (2026)

To use NotebookLM, create a notebook, upload your source documents, then ask questions, request summaries, or generate an Audio Overview — every answer is grounded in and cited back to the sources you provided.

6 min read · Updated Jul 2, 2026

To use NotebookLM, create a notebook, upload your source material, then ask questions, request a summary, or generate an Audio Overview — every response is grounded in and cited back to the specific sources you provided, rather than general AI knowledge.

Step 1: Create a notebook

  1. Sign in to NotebookLM with a Google account.
  2. Click to start a new notebook — think of each notebook as a self-contained workspace for one project, class, or topic.
  3. Name it something specific, especially if you’ll be running multiple notebooks side by side.

Step 2: Upload your sources

  1. Add your source material — PDFs, Google Docs or Slides, plain text, or web links are all typically supported.
  2. Keep sources focused on the topic at hand. NotebookLM answers only from what you provide, so uploading unrelated material dilutes the quality of grounded answers.
  3. Wait for sources to process — larger documents take a little longer to become fully searchable.

Step 3: Ask questions

Once sources are loaded, start asking questions directly:

  • Ask specific questions (“What does the report say about Q3 revenue?”) rather than vague ones (“Summarize everything”) for the most useful answers.
  • Follow up on an answer to drill deeper into a specific point, the same way you would in a normal conversation.
  • Check the citations attached to each answer — click through to verify the claim against the original source, especially for anything you plan to rely on.

Step 4: Generate summaries and study aids

Beyond direct Q&A, NotebookLM can turn your sources into other formats:

OutputWhat it’s useful for
SummaryQuick overview of long documents before a deep read
Audio OverviewPodcast-style discussion for listening on the go
Mind mapVisual layout of key concepts and how they connect

To generate an Audio Overview, select that option inside a notebook with sources already loaded — NotebookLM produces a conversational, two-voice audio discussion of the material, useful for reviewing dense content without staring at a screen.

Example: setting up a notebook for exam prep

To make the steps concrete, here’s how the workflow looks for a common use case:

  1. Create a notebook named after the course or exam.
  2. Upload the syllabus, lecture slides, and any assigned readings as separate sources.
  3. Ask NotebookLM to summarize each major topic, then ask follow-up questions on anything unclear.
  4. Generate an Audio Overview to review the material passively — while commuting, exercising, or doing other tasks.
  5. Before the exam, ask NotebookLM to generate practice questions based only on the uploaded material, and check your answers against the citations it provides.

The same pattern — create, upload, query, generate a review format, verify against citations — applies whether you’re studying for an exam, prepping for a client meeting, or working through a stack of research papers.

Step 5: Refine and reuse

  1. Add more sources to an existing notebook as a project grows, rather than starting over.
  2. Remove sources that turn out to be irrelevant to keep grounded answers focused.
  3. Share the notebook with collaborators if your workflow supports it, so a team works from the same grounded source base.

Tips for getting more useful answers

  • Curate before you upload. A notebook with 10 tightly relevant sources will consistently outperform one with 50 loosely related documents — grounded answers are only as focused as what you feed in.
  • Ask follow-ups instead of restarting. NotebookLM keeps context within a conversation, so drilling deeper on an answer usually works better than rephrasing the same broad question from scratch.
  • Use the citations as a shortcut, not a formality. Clicking through to the cited passage is often faster than re-reading the whole source when you need to double-check something specific.
  • Separate unrelated projects into separate notebooks. Mixing sources from different topics in one notebook makes grounded answers noisier and less precise.

Common pitfalls to avoid

  • Uploading low-quality scans or images of text — NotebookLM works best with clean, extractable text; poorly scanned PDFs can degrade how well it grounds answers.
  • Expecting it to know things outside your sources. If you ask a question your uploaded material doesn’t cover, don’t expect a reliable answer — that’s outside what NotebookLM is designed to do.
  • Forgetting to prune old sources. Removing material that’s no longer relevant to a notebook keeps answers focused as a project evolves.

For a fuller breakdown of what the tool is and isn’t built for, see what is NotebookLM, and check NotebookLM pricing if you’re deciding whether the free tier’s limits are enough for your project.

Beyond source-grounded research

NotebookLM is deliberately scoped to your uploaded sources — it won’t help with general coding questions, creative writing, or comparing how different AI models reason about a topic. For that broader work, Perspective AI gives you 30+ models, including GPT, Claude, and Gemini, under one pay-as-you-go subscription, so grounded research and general-purpose AI work don’t require entirely separate tools and bills.

Explore Perspective AI — every model, one subscription

Affiliate link for Perspective AI — we may earn a commission at no extra cost to you.

Frequently asked questions

How do I start using NotebookLM?

Sign in with a Google account, create a new notebook, and upload your source material — PDFs, Google Docs, web links, or text. Once sources are added, you can ask questions, request summaries, or generate an Audio Overview grounded in that content.

What file types can I upload to NotebookLM?

NotebookLM accepts PDFs, Google Docs and Slides, plain text, web links, and in some workflows audio or video files, which get processed into a searchable, citable source base for that specific notebook.

How do I generate an Audio Overview?

Inside a notebook with sources already uploaded, select the Audio Overview option and NotebookLM generates a podcast-style discussion between two AI voices summarizing your material, which you can listen to or download depending on the interface.

How do citations work in NotebookLM?

When NotebookLM answers a question, it includes citations pointing to the specific source and passage the answer came from, so you can click through and verify the claim against your original document rather than taking the answer on faith.

Can I organize multiple projects in NotebookLM?

Yes — each notebook is a separate, self-contained workspace with its own sources, so you can keep different research projects, classes, or reports organized independently instead of mixing source material across topics.

How do I get better answers from NotebookLM?

Upload focused, relevant sources rather than everything you own, ask specific questions instead of broad ones, and use follow-up questions to drill into a topic — since answers are grounded in your sources, better-curated sources produce better answers.

Can I share a NotebookLM notebook with others?

NotebookLM supports sharing notebooks with collaborators in many workflows, letting a team or study group work from the same grounded source base rather than each person uploading and querying separately.