What Is NotebookLM? Google's AI Research Assistant, Explained in Plain English

CodexFuse Team · 2026-10-27T00:07:52.638Z · 13 min read

NotebookLM is Google's AI research assistant — and if you have heard the name lately, you may also have heard its new one. In July 2026 Google renamed it Gemini Notebook, and that is the name you will see when you open it today. The tool itself did not change: you upload your own documents, and it answers questions using only those documents, with a citation pointing back to the exact passage. If you have used Claude Code or another AI helper, the idea will feel familiar — but where Claude Code writes software, NotebookLM reads your material and turns it into answers, summaries, and even a podcast. This guide explains what NotebookLM is in plain English, what it can do, what it costs, and whether your documents are safe in it.

What Is NotebookLM?

NotebookLM is Google's free, source-grounded AI research assistant. You create a notebook, upload sources — PDFs, Google Docs, web pages, YouTube videos, audio files, or pasted text — and the tool becomes an expert on just those materials. Ask it a question and it pulls the answer from your sources, not from the open internet, and every answer carries citations linking back to the exact paragraph it came from. That grounding is the whole point: because it can only answer from what you gave it, it rarely invents facts the way a general chatbot can.

Google first showed the tool as Project Tailwind at Google I/O in 2023, and it left its experimental phase in late 2024. Since then it has grown from a document chat tool into a full research workspace: it now runs on Gemini 3 and can turn your sources into audio overviews, video overviews, mind maps, flashcards, quizzes, reports, and slide decks. Students use it to study, researchers use it to work through long papers, and professionals use it to get answers from contracts, manuals, and meeting notes without re-reading everything.

NotebookLM Is Now Called Gemini Notebook

NotebookLM is now called Gemini Notebook. Google renamed it on July 16, 2026: same tool, same limits, new name and a new home at notebook.google.com. Old links redirect on their own, so a bookmark to the old address still lands in the right place. Google's reason for the change was simple — one Gemini brand across all of its AI products.

The rename matters for searchers more than for users. Most guides written before July 2026 still call it NotebookLM, and some of them describe an older, smaller tool. If you search for one name and land on a page using the other, you are reading about the same product. This article uses NotebookLM throughout because that is the name most people still search for, and notes Gemini Notebook wherever the new name is what you will actually see on screen.

One genuinely new feature arrived with the rename: a secure cloud computer in every notebook. When a question needs more than reading — adding up a spreadsheet, comparing columns, spotting a trend — the AI can write code, run it on that computer, and hand you the result. It is live today for Google AI Ultra subscribers and is rolling out to Pro users over the coming weeks. Free and Plus users do not have it yet. NotebookLM runs on Gemini 3, the same Gemini family behind Nano Banana, Google's image model — but where Nano Banana makes pictures, NotebookLM reads and reasons over your documents.

How Does NotebookLM Actually Work?

NotebookLM works in three pieces: a notebook, your sources, and a chat. A notebook is a project container with its own sources, chat, and generated outputs. A source is one item you add — a PDF, a Google Doc, a pasted block of text, a web page, an audio file, or a YouTube video. The chat is where you ask questions, and every answer is grounded in the sources you selected, with citations you can click to jump to the passage in context. Each notebook works independently: NotebookLM cannot read across two notebooks at the same time, so the way you split your notebooks decides what the AI can see.

The grounding idea is the same one behind the Model Context Protocol, which gives AI tools a standard way to work with outside data. NotebookLM applies it in the simplest possible form: it only ever reads the sources you put in the notebook. That is also why it is honest about its limits. If the answer is not in your sources, it says so rather than inventing something plausible. The trade-off is that the quality of everything it produces depends directly on the quality of the materials you upload — a thin or messy source set gives thin, messy answers.

What Can NotebookLM Do?

Once your sources are in, NotebookLM does far more than answer questions. Its Studio panel turns the same material into different formats, so the same research paper can become a podcast for the drive, a study guide for the exam, or a slide deck for the meeting. Here is what it handles, all derived from your uploaded content.

The Audio Overview is the feature most people know NotebookLM for, and it is worth the reputation. Two hosts discuss your material in a conversational way, which makes a dry paper or a long report easy to absorb while walking or driving. You can customize it before it generates — ask it to go deep on one topic, or to explain like you are five — and the interactive mode lets you interrupt the hosts with a question in real time. It is still in beta, so expect the occasional rough edge.

How to Use NotebookLM in 5 Steps

Getting started takes about five minutes. You need a Google account and at least one source — a PDF, a web page, or even a pasted block of text will do.

1. Go to notebook.google.com and sign in. That is the tool's new home under the Gemini Notebook name. Your old NotebookLM links redirect there automatically.
2. Create a notebook. Click Create New Notebook and give it a name that matches the topic or project, because each notebook is its own sealed container.
3. Add your sources. Upload PDFs, Docs, slides, web pages, YouTube links, or audio files — or use Discover Sources and type your topic to let NotebookLM find sources for you. Check the list and remove anything weak before you start asking questions.
4. Read the summary, then ask. NotebookLM generates an automatic summary of your sources. From there, ask questions in the chat, or pick one of the suggested questions it offers based on the content.
5. Generate a Studio output. In the Studio panel, pick what fits the job: an Audio Overview to listen to, a study guide to revise from, a mind map to see the structure, or a briefing doc to share.

One habit makes the difference between good answers and great ones: how you ask. The quality of what comes back depends on the question, so learning to write better AI prompts pays off here just as it does with any other AI tool. Ask for specifics — "summarize the three arguments in section 2, with a citation for each" — and you will get specifics back.

Is NotebookLM Free? Plans and Prices

NotebookLM is free to use, and the free plan is enough for most people. It gives you 100 notebooks, 50 sources per notebook, and a standard usage budget that refreshes every 5 hours, with audio and video overviews included. The paid options are bundled into Google AI plans, and the table below shows the tiers as listed by Elephas, checked on October 27, 2026. Google changes these plans, so treat the table as a snapshot and check the current page before you pay for anything.

There are also daily caps on the free plan: 50 chat queries per day and 3 audio generations per day, according to Carnegie Mellon University's guide for its students and staff. Each source can hold up to 500,000 words or 200 MB, and that ceiling is the same on every tier — paying more buys you more notebooks, more sources, and a bigger usage budget, not bigger individual files. For most students and casual researchers, the free limits are generous. The paid tiers make sense when you run many notebooks at once or generate audio and video overviews every day.

NotebookLM vs ChatGPT: What's the Difference?

The difference is where the answer comes from. ChatGPT draws on a broad pre-trained dataset of the internet and is built for open-ended conversation about almost anything. NotebookLM draws only on the sources you upload and is built for focused work on those sources. That one difference decides which tool to reach for.

The two tools are complements, not rivals. Use ChatGPT when you need ideas, drafts, or an answer about the world. Use NotebookLM when you need an answer that must be true to a specific set of documents — a paper you have to cite, a contract you have to understand, or a report you have to summarize accurately. If your question is about the live web rather than your own documents, an AI browser that searches and reads pages for you is the better tool for that job.

Is NotebookLM Safe? What Happens to Your Documents?

Your files are processed on Google's servers, not on your device, so NotebookLM is not the best fit for confidential material. That is the honest answer, and it is the one every guide that has actually used the tool gives. Google says that for individual users, your data is generally not used for training unless you explicitly choose to share feedback with Google. For organizations and schools using Workspace versions, data stays private within their environment and is not used for model training — Carnegie Mellon University's guide confirms that when students log in with a university account, Google does not use their prompts and responses for training, and uploaded documents are not retained beyond the session.

Three practical rules follow from that. First, do not upload anything you would not want sitting on Google's servers — contracts with sensitive terms, medical records, or unpublished work with real commercial value. Second, remember that a notebook is sealed: what you upload to one notebook cannot be read from another, which limits accidental exposure but also means you should not duplicate sensitive files across notebooks. Third, web pages and YouTube videos you add are not updated automatically, so re-check a source if the original page has changed since you added it. NotebookLM can still occasionally give a false output, so double-check anything important against the citation it points to — the citation is there precisely so you can.

Who Should Use NotebookLM — and Who Should Skip It?

NotebookLM is best for students, researchers, and professionals who work with documents and need trustworthy answers from them. A student can turn a semester of readings into flashcards and a podcast. A researcher can ask a 300-page report for the three findings that matter, with citations. A lawyer or analyst can work through a contract or a set of meeting notes and get a briefing document instead of re-reading everything. Anyone who wants to learn a topic faster, with the sources in front of them, will get value from it — and the free plan covers that use comfortably.

Skip it, or use it carefully, in three cases. If your documents are confidential, the server-side processing is a real drawback and a local, on-device tool fits better. If you need answers about the wider world rather than your own files, a general chatbot or an AI browser is the right tool. And if you need several notebooks to talk to each other, NotebookLM cannot do that yet — each notebook stays independent. For everyone else, it is the rare AI tool whose main promise is restraint: it only knows what you gave it, and it tells you where every answer came from.

Sources

• Elephas, "What Is Gemini Notebook (NotebookLM)? Features & Pricing 2026" (updated October 2026) — the July 16, 2026 rename to Gemini Notebook and the notebook.google.com home, Gemini 3 engine, Audio and Video Overviews, mind maps, flashcards, quizzes, reports, Deep Research, the secure cloud computer and its Ultra/Pro availability, free limits (100 notebooks, 50 sources per notebook, 500,000 words or 200 MB per source, usage budget refreshing every 5 hours), the pricing table (Free $0, Plus $4.99, Pro $19.99, Ultra $99.99 and $199.99), and the server-side processing privacy warning. https://elephas.app/blog/what-is-notebooklm
• DigitalOcean, "What Is NotebookLM? Features and How to Use It in 2026" — the source-grounded FAQ: what NotebookLM is used for, the free tier with daily and monthly limits, the NotebookLM vs ChatGPT difference (closed retrieval system, fewer hallucinations, citations), and Google's training-data policy for individual versus Enterprise users. https://www.digitalocean.com/resources/articles/what-is-notebooklm
• Carnegie Mellon University, "How to Use NotebookLM" — the free-tier limits (100 notebooks, 50 sources and 500,000 words each, 50 chat queries and 3 audio generations per day), the upload types (Google Docs, Slides, PDFs, web URLs, YouTube videos, audio files), and the Workspace for Education protection (prompts not used for training, documents not retained beyond the session, FERPA-compliant use). http://www.cmu.edu/computing/services/ai/tools/google-ai-tools/notebook-lm-how-to.html
• American Bar Association, "NotebookLM: Turning Documents into a Research Assistant" (Law Practice Magazine, March–April 2026) — the history (Project Tailwind at Google I/O 2023, experimental phase ending late 2024) and the source-grounded value for professional research. https://www.americanbar.org/groups/law_practice/resources/law-practice-magazine/2026/march-april-2026/notebooklm-turning-documents-into-a-research-assistant/
• YourTechCompass, "NotebookLM: Google's AI Research Tool That Reads Your Sources" — the Gemini 3 engine, Audio Overviews, Video Overviews, Cinematic Video Overviews, Mind Maps, Infographics, Slide Decks with PPTX export, Flashcards, Quizzes, Data Tables, and Deep Research reports. https://yourtechcompass.com/notebooklm-review/
• BillionVerify, "NotebookLM: A Complete Guide for Marketing and Sales Teams" — the notebook, source, and chat structure, and the Project Tailwind origin. https://billionverify.com/blog/notebooklm

Frequently Asked Questions

What is NotebookLM in simple terms?

NotebookLM is Google's AI research assistant that answers questions using only the documents, web pages, videos, and audio files you upload. You create a notebook, add sources, and ask questions — the tool answers from those sources only, with citations pointing to the exact passage, so it rarely makes things up.

Is NotebookLM now called Gemini Notebook?

Yes. Google renamed NotebookLM to Gemini Notebook on July 16, 2026. It is the same tool with the same limits, just a new name and a new home at notebook.google.com. Old links redirect automatically. This article uses NotebookLM because that is still the name most people search for.

Is NotebookLM free?

Yes, the free plan is enough for most people. It includes 100 notebooks, 50 sources per notebook, a usage budget that refreshes every 5 hours, and Audio and Video Overviews. Daily caps apply: 50 chat queries and 3 audio generations per day. Paid tiers start at $4.99 a month for Plus and go up to $199.99 for Ultra, bundled into Google AI plans.

How is NotebookLM different from ChatGPT?

ChatGPT answers from broad pre-trained knowledge of the internet, while NotebookLM answers only from the sources you upload. NotebookLM gives inline citations to the exact passage, makes things up far less often, and can turn your documents into podcasts, videos, mind maps, and study guides. ChatGPT is better for open-ended questions, brainstorming, and writing.

Is NotebookLM safe for confidential documents?

Not really — your files are processed on Google's servers, not on your device. Google says personal data is generally not used for training unless you share feedback, and Workspace for Education accounts get stronger protection, but you should not upload contracts, medical records, or other sensitive material. For confidential files, an on-device tool is the safer choice.