Notebooklm

Gemini Notebook is Google's source-grounded research tool. Until July 16, 2026 it was called NotebookLM, and most people still know it by that name; the product is the same, with the old notebooklm.google address redirecting. You create a notebook, add the sources that matter, and then ask questions, get summaries, and generate outputs that draw only on those sources, with inline citations that point back to the exact passage.

It is aimed at anyone who has to understand a pile of documents rather than the whole internet: a founder working through a data room or a set of customer interviews, a marketer digesting a research report, an analyst who needs the numbers from twelve PDFs in one place. Because it will not answer from outside your sources, it behaves less like a chatbot and more like a research assistant who has only read what you handed over, which is exactly what you want for accuracy. It is the practical face of grounding and RAG for non-developers.

Highlights

• Source-grounded answers with citations — chat with the notebook and every answer cites the source passage it came from, so checking is one click.
• Broad source support — Google Docs, Slides, Sheets, PDFs, Word, PowerPoint, Markdown, CSV, ePub, images, audio, web pages, public YouTube videos, pasted text, and Gemini chats, up to 500,000 words or 200 MB per source.
• Audio and Video Overviews — generate a podcast-style discussion or a narrated video explainer of your sources; audio can be downloaded for offline listening in the mobile app.
• Study and briefing outputs — mind maps, flashcards, quizzes, infographics, slide decks, and reports built from the same sources.
• Code execution over your sources — since the July 2026 rename, each notebook has a secure cloud computer that can write and run code for data analysis grounded in your sources, rolling out first to Google AI Ultra and Workspace, then Pro.
• Inside the Gemini app — notebooks now appear in the Gemini app with cross-app syncing, and Google has said they are coming to AI Mode in Search.

In a founder or analyst workflow

The pattern is one notebook per decision. Load everything relevant, then ask questions that force the notebook to reconcile sources against each other rather than summarize each one:

Sources: our last 3 board decks, the Q2 P&L export, and the two analyst
reports on the category. Where do the analyst growth assumptions differ
from the numbers in our own decks? Quote the specific figures from each
source and flag any place a deck contradicts the P&L.

When the answer holds up, generate an Audio Overview for the drive to the meeting, or a slide deck as the skeleton for the real one. For the write-up itself, hand the cited findings to Claude or ChatGPT, which are stronger at long-form drafting; Notebook is for understanding, not for prose.

Good to know

Gemini Notebook runs on the web and in Android (Android 10 and up) and iPhone and iPad apps, with iOS availability more limited by country. Plans as of September 2026, per Google's upgrade page: Free with a Google Account (100 notebooks, 50 sources per notebook, 50 chats and 3 Audio Overviews a day), Google AI Pro at $19.99 per month (500 notebooks, 300 sources, 500 chats and 20 Audio Overviews a day, plus higher access to Gemini models), and Google AI Ultra from $99.99 per month (600 sources per notebook, 5,000 chats and 200 Audio Overviews a day). Paid plans also remove the watermark on generated media in most countries.

With a work or school account, Gemini Notebook is a core Workspace service in Business Starter, Business Standard, Business Plus, and the Enterprise and Education editions, with the enterprise data protections those plans carry: uploads, chats, and outputs are not reviewed by humans or used to train models. Google Cloud also sells Gemini Notebook for Enterprise with higher artifact limits and controls such as VPC Service Controls and IAM. Google says more than 30 million people and 600,000 organizations use the product.

If your questions need the open web rather than a fixed document set, Perplexity is the closer tool; if you need the documents acted on rather than explained, look at Claude.