How to Use Gemini Notebook Better than 99% of People

Parker Prompts· 11 min· 2,601 words· 12 min read· English ·Watch on YouTube

This is the full transcript of How to Use Gemini Notebook Better than 99% of People, published on YouTube by Parker Prompts. Every paragraph carries the moment it was spoken, so you can click any line to jump straight to that point in the video, search the whole thing for a word, or copy it out.

0:00Whether you're a professional, a student, or just someone trying to learn faster and think clearer in 2026, you already know about Notebook LM. Notebook LM, now Gemini Notebook, is one of the best tools Google has made, but most people only use 10% of it. So, in this video, I'll show you the other 90% [music] and by the end, you'll have a notebook that finds its own sources and keeps itself current without you. So, here on the home screen, I'll click create new notebook and the first thing I'll do is rename it to match what I'm researching, which for me is machine learning. The notebook takes that title

0:27as the subject it's working on, so everything it gives you is built around it. It keeps you organized as well and with 100 notebooks on the free plan, the name is how you find this one again. Once you're inside, the screen splits into three with your sources down the left, the chat in the middle, and the studio panel over on the right. Those three panels are the whole product and the left one sets how good the other two are. The notebook only answers out of what's in that sources panel and that's why it works nothing like asking ChatGPT or Claude the same question. You hand it

0:52your material first and every answer afterwards is built out of that material and nothing else with a citation pointing back at the line it came from. So, the notebook's named and pointed at your topic and that sources panel is the only thing left to fill. And you don't have to do that manually because the next thing I'll show you does it for you. In that sources panel, there's a search box and it pulls sources in without you going and finding them. You type in whatever you're researching and choose web or drive. Then it comes back with a list you go through yourself before anything gets imported. That one

1:19is fast research and it's been in there since November 2025. Searching drive works in plain English, too. So, I'll type slides from Claire, Google's own example, and it finds them without me knowing the file name. Next [music] to it is deep research and that's the one to use when you're starting a topic from nothing at all. Turn web and deep research on together and it reads up to hundreds of websites, then builds a multi-page report with the sources behind it. It runs in its own time while you carry on and results come back a couple of minutes later. Once the results are up, every result has an

1:46import and a delete button next to it. And if you minimize the panel, the list is still there when you come back to it. One thing while we're on this screen, the name at the top says Gemini Notebook because Google renamed it on the 16th of July. Every shared notebook and every link still works because it redirects automatically. So, it's only the name

2:02and the logo you have to get used to. And that's a notebook full of sources without opening a single tab to collect them. That covers what it finds for you, and the rest of that panel is everything you can hand at yourself. The list of what it actually takes is longer than you'd think, and a few things on it surprised me. The YouTube one is where I'd start if you're studying because you paste in the link to a public lecture with captions, and its transcript becomes a source you can question like any other. So, a 2-hour lecture on derivatives goes in as a link with no

2:26limit on length unless the captions run past 500,000 words, and you can ask about the chain rule and get the transcript quoted straight back. Same if you're writing a paper because a conference talk or an interview works the same way as long as it's public and captioned, and you pull quotes straight out of the transcript. EPUB came much later in March this year, so you can drop an entire ebook in as one source and never break it into chapters. Audio files like MP3 and WAV go in, images in 14 different formats, up to 100 slides at once, and sheets up to 100,000 tokens. Word documents, plain text,

2:56markdown, PDFs, CSVs, and PowerPoint all go in. And so do web URLs and anything you paste straight into the box. Gemini chats count as a source, too, and that one starts over in the app itself because you open the chat there and pick add to notebook. Once you've done that, the chats you have with that notebook over in the Gemini app turn up as context in your sources panel. On your

3:15phone, it all works off the share icon. So, if you're on a website, a PDF, or a YouTube video, you tap share, pick Gemini notebook, and it goes in as a new source on a notebook you already have. Each source holds up to 500,000 words. So, the size cap is about the only thing

3:28left that stops something getting in. Getting everything in is the easy part, though, because a notebook holding 50 sources gives you vaguer answers than one holding three. Every source has a checkbox next to it, and ticking the wrong ones is the whole reason an answer comes back vague. Untick a source, and it comes straight out of the answer, though it stays in the notebook so you can put it back whenever you want. Once you have five or more sources, it labels and categorizes them by a on its own, and you can rename those labels, add your own, delete them, or move sources between them. Anything you brought in

3:56from Drive syncs itself every few minutes, too. So, when the doc changes, the notebook changes with it. Citations are still in line through all of this. So, hover over one and you get the quoted text, click it, and you jump to the exact spot in the source. Images can be citations as well. And if a source is really short, it sites the whole document instead of picking out a bit of it. So, when an answer comes back too general, untick everything you're not asking about and run it again. That's the reading side handled. And one button in the chat panel decides how it talks to you. That button is called configure

4:23chat, and two [music] settings behind it change every answer you get. Conversational style comes first, and default is what you're already on, which Google set up for general research and brainstorming. [music] Switch to learning guide when you're trying to learn the material properly because it works through the ideas with you as you go. Custom lets you type your own, so you can tell it to answer like a PhD student or to run the whole notebook

4:42like the host of a role-playing game. Response length is right under it, and default, longer, and shorter change the answer more than the style does. Chat history saves itself and stays private to you even inside a shared notebook. So, you can close it and pick it up later. When an answer is worth keeping, save to note pins it to the note board with the tables and the clickable citations still attached. Both of those settings stay until you change them, so you set them once for each notebook and then forget about them. Now, the panel on the right is where the notebook builds the stuff a chatbot can't. I use

5:10two things in the studio panel way more than everything else on that list put together. The full list is audio overviews, video overviews, mind maps, [music] reports, data tables, flash cards, or quizzes, slide decks, and infographics with an add note button below them. Data tables pulls your sources into a table you can send straight to sheets, and the citations go on the second tab. So, I'll ask it to pull every number out of the six papers I've loaded, and I get back a table I can sort and filter. And as you can see, the columns are already labeled, and export to sheets in the three dots menu drops the whole thing

5:41into a new sheet. The other one is add note, and it's the one thing in that panel that turns back into a source. Anything you keep with save to Note goes there, and you open a note and hit convert to source for one, or use more in the note section and convert all notes to source for the lot. A saved answer becomes a source the next question can read, so the notebook gets better the longer you use it. And every one of those outputs has its own formats

6:01behind it, starting with the audio one. Deep Dive is the two-host podcast you already know, and the three formats next to it do something else entirely. [music] The brief is one host giving you the key points on their own, and it's over in under 2 minutes. Run the critique on something you wrote, and it tells you which parts are weak. And the debate takes the same material and argues both sides of it. All four come in more than 80 languages, [music] and in English you can make them shorter or longer. So I'll hit join and ask them how to explain fractional calculus to my 60-year-old mom who knows nothing about

6:30AI, and they stop and answer me before carrying on. Scientists call it truncation window. It literally caps that mathematical memory at a finite length. So it gives the algorithm the benefit of historical context, but without the crushing computational weight. Exactly. Oh, our listeners got something share. Let's hear it. Hey, could you please explain fractional calculus to my mom? She doesn't know

6:49much about AI. I love that. Let's make this something she could relate to over a cup of coffee. Think about it like this. Standard calculus is like looking at a single photo of a moving car. Right. It'd tell you the speed in that exact moment, but nothing about where it came from. Exactly. Now tell her fractional calculus is like a video that looks back

7:06at the last mile. So instead of just seeing the speed now, you're seeing the whole journey and the traffic patterns. It's using a fading memory of the past to make a better decision about the next turn. Video overviews split into cinematic, explainer, and short. An explainer is the only one with all the languages,

7:21while the other two are English only. Explainer is also the one with visual style on it, so you get classic, whiteboard, watercolor, retro print, heritage, paper craft, kawaii, and anime, or you let it pick, or you describe a custom look yourself. [music] Reports let you choose blog post, study guide, or briefing document, or an AI suggested report type, and the pencil icon is where you set your own structure and tone before you generate. And flashcards, quizzes, and infographics all take a difficulty or detail setting plus a prompt. So, point them at the one chapter you're weak on and not the whole notebook. On a mind map, you click a

7:53single node and ask a question about that one branch, so you're not scrolling around a huge map to find it. Five of these are over 18s only. Slide decks, infographics, the cinematic and short video formats, and deep research from earlier. The format you pick decides more about what you get than anything you type into it. And every one of them comes out of an allowance you can't see until you hit it. The free plan is called Gemini Notebooks Standard, and it gives you three audio overviews and three video overviews a day, 50 chats, and 10 reports, flashcards, quizzes, and mind maps. Deep research is the tightest

8:21one. 10 a month on free, three a day on plus, and 20 a day on pro. Slide deck edits come out of that same allowance as well. So, put every change you want into one request. Everything resets on a rolling 24 hours, and the monthly ones after 30 days. So, using them all up

8:35today doesn't leave you short all week. Everything so far has been inside that one tab, and your notebook stops staying in it this year. Since April, they're in the Gemini app as well. And any source you add in one place turns up in the other. That one is personal Google accounts only, so work and school accounts don't get it yet. Shared notebooks never show up in Gemini even when you own them, and a notebook with a Gemini chat in it can't be shared at all. Your notebooks also show up in AI mode in Google search, and there are the answers use your own sources on top of

9:00the web. Then there's Workspace Studio, and there you don't open the notebook at all. It just runs as one step inside an automation. That's everywhere the notebook shows up now, but Google also changed what it can actually do once you're inside it. Ask it for a chart, and it writes the code and runs it for you using whatever sources you've loaded. Google put a secure cloud computer inside every notebook in June, and it moved on to Gemini 3.5 at the same time, which is when you started

9:23being able to expand its thinking steps. What comes out of it is proper files. So, ask for a chart, and you get a PNG or an SVG. Ask for a budget, and you get an Excel file, PDFs, Word documents, Markdown, CSVs, JSON, and PowerPoint all come out the same way, and you download them straight from the studio panel. So, I'll ask for one chart of the spend across all six sources, and the PNG is in studio before I finished reading the answer. The chart comes back with the access labels taken straight off the source data, and I haven't typed a thing. On Pro and Ultra, the chat will

9:53search the web, run code, and build those files for you, whether you've got sources loaded or not. It started on Ultra and Workspace accounts with expanded AI access, and it's on Pro now, too, though desktop only. Any of this can go to someone else in two clicks. Share is in the top right corner. Only owners and editors can make a public link, and a globe icon next to it tells you when a notebook is public. There's a copy link to chat view option that gives someone just the chat, and Google says outright that it hides the material without taking anyone's access away. So, don't ever treat that one as a privacy

10:22setting. Every setting and every output in that notebook is one you now know how to use. And the one thing that notebook can't do is hold a face study across more than one shot. Consistency is the whole problem with AI video, and there's a system that solves it across two tools. [music] So, I made a video on how to do exactly that with Gemini and Higgs

10:38Field, and you can watch that one next. Thank you for watching, and I'll see you in the next one.

Where these words come from. This is the caption track YouTube holds for this video, written automatically by YouTube rather than by the creator. We read it, tidied the line breaks and laid it out so it can be read. The plain text version is at https://viewrankai.com/tools/youtube-transcript/kxi3bGvUnLQ.txt.

All rights in this video belong to Parker Prompts. Watch it on YouTube. If this is your video and you would rather this page did not exist, tell us and we will remove it.