Skip to content

Gemini Notebook (formerly NotebookLM)

Google's source-grounded research notebook: upload documents, then chat, get Audio Overviews, reports and quizzes from them.

Reviewed by Andrew Miller · Updated

Our rating

4.6/5

For working from your own documents, Gemini Notebook is the strongest free tool going, and its opaque new compute limits are the price of that generosity.

What it is

Gemini Notebook is Google's research tool that answers only from the sources you give it: PDFs, Google Docs, websites, YouTube videos and more. It was called NotebookLM until July 16, 2026, when Google renamed it; the product itself did not change, and the old notebooklm.google.com address redirects to notebook.google.com.

Beyond chat, it can turn sources into Audio Overviews (podcast-style discussions), Video Overviews, reports, mind maps, flashcards, quizzes, infographics and slide decks. In 2026 Google began adding a secure cloud computer to each notebook, so it can write and run code to analyse data, launched first for AI Ultra and coming to Pro. Notebooks also open in the Gemini app and sync both ways.

It is free, with higher limits on Google AI plans, and available to Workspace and Gemini Enterprise customers. Google says more than 30 million people and 600,000 organisations use it.

Our review

Gemini Notebook's value is discipline. It answers from the sources you give it, with citations back to the passage, and does not wander onto the open web. That makes it more trustworthy than a general chatbot for course readings, a stack of research papers or a due-diligence data room. Google says more than 30 million people and 600,000 organisations use it, and the July 2026 rename from NotebookLM changed the name and the address, not the product.

The free tier is generous on size: 100 notebooks of up to 50 sources each. AI Plus at $4.99 a month doubles both to 200 and 100, and AI Pro at $19.99 allows 500 notebooks of 300 sources. Since 2 September, though, usage is no longer counted per feature per day. A compute budget refreshes every five hours up to a weekly cap, shared across chat, Audio and Video Overviews, reports and Deep Research, and Google publishes only multipliers (Plus 2x, Pro 4x, Ultra 5x or 20x Pro) rather than the underlying numbers. A student cannot know in advance how many Video Overviews a week will allow.

The privacy terms split by account. On a personal Google account, Google says your data is not used to train the model unless you send feedback, in which case reviewers may see the whole interaction, uploads included. On Workspace editions where it is a core service, which include Business Starter and above, Enterprise and Education, uploads, queries and answers are not reviewed by people or used for training. Google's admin help adds that Workspace data-region settings do not apply to data the tool processes and caches, a detail that matters to an EU firm with a residency clause.

The 2026 additions push it past summarising. Each notebook is getting a cloud computer that writes and runs code over your sources, on AI Ultra first and coming to Pro, and Gemini Enterprise customers can keep files inside their own Google Cloud project behind VPC Service Controls. The visible watermark on generated images and video can be switched off only on Pro and Ultra, and an invisible SynthID watermark is always applied.

It remains weak at open web research, where Perplexity and Gemini's own Deep Research do better. Its strength is the opposite habit: refusing to know more than the pile of documents in front of it.

Where it earns its place

  • A second-year law student

    Revise from 40 case readings before exams

    Load the readings into one free notebook (50-source limit), generate flashcards and quizzes, and check each answer's citation. Heavy Video Overview use can exhaust the five-hour budget; eligible US students can get AI Pro free for a year if they redeem by 31 December 2026.

  • An analyst at a 50-person consultancy on Google Workspace Business Standard

    Query a client's document pack without it being used for training

    On Business Standard Gemini Notebook is a core service, so uploads are not reviewed by people or used to train models; the firm should still note that data-region settings do not extend to the tool's processing and cache.

  • A data platform engineer at a bank running on Google Cloud

    Give analysts notebooks over internal reports without files leaving the cloud project

    Gemini Notebook for Enterprise, sold with Gemini Enterprise plans, keeps files in the customer's Google Cloud project with IAM and VPC Service Controls; pricing goes through Google Cloud and is not on the plans page.

Before you commit

Setup
Sign in with a Google account at notebook.google.com, add PDFs, Docs, web pages or YouTube links, and get a cited answer within minutes; notebooks sync with the Gemini app.
Integrations
Google Drive, Docs and Slides as sources; notebooks open in the Gemini app and are coming to AI Mode in Search; the Enterprise version runs inside Google Cloud.
Data and privacy
Personal accounts: not used for training unless you give feedback, which reviewers may read. Qualifying Workspace and Education editions: no human review and no training. Enterprise keeps files in your Google Cloud project.
Cost at scale
Free for most individuals; 20 people on individual AI Pro plans pay about $400 a month, while Workspace customers get it inside existing licences at limits set by edition.
Limits
Compute budget refreshes every five hours up to a weekly cap; free tier 100 notebooks and 50 sources per notebook, Pro 500 and 300, Ultra up to 600 sources; Google marks all limits as subject to change.

What's good

  • Answers grounded in your sources, with passage-level citations
  • Free tier allows 100 notebooks of 50 sources each
  • Qualifying Workspace editions: no human review or training on your data
  • Code execution over sources on higher plans

What's not

  • Compute-based limits with no published numbers
  • Sending feedback on a personal account opens the whole exchange to reviewers
  • Visible watermark can only be turned off on Pro and Ultra
  • Workspace data-region settings do not cover its processing and cache
  • Not built for open web research

Best for

Students, researchers and analysts who need to understand and cite a defined set of their own documents.

Alternatives

Gemini Notebook (formerly NotebookLM): common questions

Is Gemini Notebook (formerly NotebookLM) worth it?
For working from your own documents, Gemini Notebook is the strongest free tool going, and its opaque new compute limits are the price of that generosity. We rate it 4.6 out of 5.
How much does Gemini Notebook (formerly NotebookLM) cost?
Free; higher limits through Google AI Plus ($4.99/month, 2x), AI Pro ($19.99/month, 4x) and AI Ultra (about $100 or $200/month, 5x or 20x Pro). Included in qualifying Workspace editions; Enterprise version through Google Cloud.
What are the downsides of Gemini Notebook (formerly NotebookLM)?
Compute-based limits with no published numbers; sending feedback on a personal account opens the whole exchange to reviewers; visible watermark can only be turned off on Pro and Ultra; workspace data-region settings do not cover its processing and cache; not built for open web research.
Who is Gemini Notebook (formerly NotebookLM) best for?
Students, researchers and analysts who need to understand and cite a defined set of their own documents.
How much does Gemini Notebook (formerly NotebookLM) cost for a team or in production?
Free for most individuals; 20 people on individual AI Pro plans pay about $400 a month, while Workspace customers get it inside existing licences at limits set by edition.
Is Gemini Notebook (formerly NotebookLM) safe for company data?
Personal accounts: not used for training unless you give feedback, which reviewers may read. Qualifying Workspace and Education editions: no human review and no training. Enterprise keeps files in your Google Cloud project.
What are the best Gemini Notebook (formerly NotebookLM) alternatives?
The closest alternatives we have reviewed are Perplexity, Elicit and Google Gemini.

Answers checked Oct 23, 2026.

Make Gemini Notebook (formerly NotebookLM)? Show this rating on your siteGemini Notebook (formerly NotebookLM) rated 4.6 out of 5 on ReviewRamble