The app-switching is finally over.
If you've used NotebookLM seriously -- for research, studying, document analysis, anything -- you know the workflow: open Gemini to start a conversation, realize you need your notebook context, switch tabs to NotebookLM, copy something back, switch again. Annoying in a low-stakes way, but annoying.
Google fixed it on April 8, 2026. Notebooks are now built directly into the Gemini app, with bidirectional sync to NotebookLM. Whatever you add in Gemini shows up in NotebookLM. Whatever you add in NotebookLM shows up in Gemini. One knowledge base, two interfaces.
That's the short version. The longer version has some nuance worth knowing before you dive in.
What Is the Gemini Notebooks Feature?
Think of it as a persistent, source-grounded workspace that lives inside the Gemini app.
Regular Gemini conversations are stateless -- great for quick questions, not great for deep projects where you need Gemini to stay anchored to your specific documents. Notebooks fix that. You upload your sources once, and then every conversation inside that notebook has access to them. Gemini references your material rather than falling back on its general training data.
The sync with NotebookLM is the clever part. NotebookLM has always been excellent at this kind of source-grounded work -- it's the whole product's reason for existing. What it lacked was integration with Gemini's broader capabilities: longer conversations, custom instructions, the ability to weave in Gemini's general knowledge when you need it. Now they work together.
Practically: you might add your research papers to a notebook in Gemini, have a long analytical conversation, then flip over to NotebookLM to generate an Audio Overview that turns those papers into a podcast discussion. Same sources, no re-uploading.
Who Can Access Notebooks Right Now
Quick reality check before you go hunting for the feature: it's not available to everyone yet.
The April 8-10 rollout went to Google AI Ultra, Pro, and Plus subscribers on web. If you pay for Gemini Advanced (which is part of Google One AI Premium at $19.99/month), you're in. Free Gemini users will get access eventually -- Google said "coming weeks" -- but no specific date.
Also not available at launch: Google Workspace accounts, Education accounts, accounts for users under 18, and mobile users. Mobile is coming soon. The Workspace/Education exclusion is more significant and Google hasn't given a timeline.
So if you're on Gemini Advanced and using the web app, you should have it. If you don't see it yet, give it a few days -- the rollout was gradual.
Step-by-Step: Creating Your First Notebook
OK so here's what the actual experience looks like.
Step 1: Go to gemini.google.com. You need the web app for now -- the mobile app doesn't have Notebooks yet.
Step 2: Open the menu. Click the hamburger icon (three horizontal lines) at the top left of the sidebar. You'll see a Notebooks section below your recent conversations.
Step 3: Create a new notebook. Click "New notebook." Give it a name that'll make sense to you in both Gemini and NotebookLM -- "Q2 Competitor Research" or "Machine Learning Fundamentals," whatever fits your project.
Step 4: Add your sources. This is where it gets interesting. Click Add sources and you'll see options for:
- PDFs -- Upload directly, or add via URL
- Website URLs -- Paste in any public web page
- YouTube videos -- Public videos only, and they need captions (auto-generated is fine). Gemini imports the transcript, not the video itself
- Google Docs -- Direct integration; link rather than export
- Copy-pasted text -- For things that don't fit neatly into a file format
You can add up to 600 sources depending on your plan. In practice, most notebooks will have a handful of focused sources rather than hundreds -- but it's good to know the ceiling isn't a problem.
Step 5: Start chatting. Once your sources are loaded, just start asking questions. "Summarize the main arguments across these papers." "What are the contradictions between these sources?" "Draft an outline for a report on this topic using only these documents."
Gemini will stay grounded in what you've uploaded. If something isn't in your sources, it'll tell you -- it's not going to silently fill gaps with hallucinated information.
Step 6: Switch to NotebookLM when you need its specific features. Open NotebookLM and your notebook is already there. No importing, no setup. Use it for Audio Overviews (the podcast-style summaries), Infographics, or the kind of deep citation-tracking that NotebookLM has always done better than anything else.
What You Can Actually Do With This
Research synthesis is the obvious use case, but the integration opens up some combinations that weren't really possible before.
Research and writing workflows. Add your source material to a Gemini notebook, have an analytical conversation to develop your thinking, then ask Gemini to draft a document structure -- all with the same sources in scope. No more "I need to reference that paper I was looking at earlier" moment.
Study guides and exam prep. Load up your course readings, lecture notes, and any relevant papers. Ask Gemini to quiz you, identify gaps, explain concepts, or generate practice questions. Then flip to NotebookLM for an Audio Overview of the material -- weirdly effective for retention while commuting.
Document analysis. Legal contracts, financial reports, technical specifications -- upload them and have an actual conversation about what they mean rather than just getting a static summary. The persistent context means you can ask follow-up questions without re-explaining what you're working with.
Audio Overviews (still in NotebookLM, but now easier to reach). This is the feature that made a lot of people take NotebookLM seriously when it launched. You upload your sources, and NotebookLM generates a podcast-style dialogue between two AI hosts discussing your material. It's surprisingly good for dense content. With bidirectional sync, the friction to get there dropped considerably.
Gemini Notebooks vs. ChatGPT Projects
The obvious comparison. Both create persistent workspaces tied to an AI assistant. They're solving the same problem differently.
| Feature | Gemini Notebooks | ChatGPT Projects |
|---|---|---|
| Source types | PDFs, URLs, YouTube, Google Docs, text | PDFs, images, files (no YouTube/URLs natively) |
| Max sources | Up to 600 | No published limit |
| Bidirectional sync | Yes -- syncs with NotebookLM | No external sync |
| Audio Overviews | Yes (via NotebookLM) | No |
| Grounding approach | Strictly source-anchored | Blends sources + general knowledge |
| Custom instructions | Yes | Yes |
| Price | Gemini Advanced ($19.99/mo) | ChatGPT Plus ($20/mo) |
| Mobile | Coming soon | Available now |
The real difference is philosophy. ChatGPT Projects blends your uploaded content with OpenAI's broader knowledge base -- useful for creative work, potentially messy for research where you need to know exactly where an answer came from. Gemini Notebooks, following NotebookLM's design, keeps answers strictly grounded in your sources. If it's not in your documents, it says so.
For researchers, students, and anyone doing serious document analysis: Gemini Notebooks is the stronger choice. For general productivity and creative projects: ChatGPT Projects has the broader ecosystem.
Our full AI tool comparisons break down more of these head-to-heads if you want to dig further.
Tips and Limitations
A few things worth knowing before you get into it:
YouTube videos need captions. Not all videos have them, and if they don't, you can't add the video. Auto-generated captions count -- most major creators have them -- but obscure content or non-English video without subtitles won't work.
Mobile isn't here yet. If you're used to doing research work on your phone or tablet, you'll have to wait. Web-only at launch.
Workspace accounts are excluded. This one caught a lot of people off guard. If your primary Google account is a Google Workspace account (company or school email), you don't have access yet. Google is working on Workspace availability but hasn't given a date.
It's not a full replacement for NotebookLM. Audio Overviews, Infographics, and some of NotebookLM's more specialized citation tools are still only in NotebookLM. The integration means you can access them, but you still need to open NotebookLM to use them.
The 600-source limit is generous but not infinite. For most research projects, you'll never get close. But if you're trying to ingest an entire academic field, there are limits to what a single notebook can hold.
Custom instructions matter. Don't just dump sources in and start chatting. Spend 30 seconds setting a custom instruction -- something like "I'm writing a report for a non-technical audience, keep explanations accessible" or "Focus only on methodology sections, not conclusions." It shapes every response in the notebook.
Should You Upgrade to Gemini Advanced for This?
Honest answer: if Notebooks is the only reason you're considering Gemini Advanced, probably not yet. The feature is solid but still rolling out, mobile isn't ready, and Workspace users are locked out.
But if you're already using NotebookLM seriously, already on Gemini, or doing the kind of work where persistent, source-grounded AI conversations would replace multiple tools in your workflow -- then yes. The integration is genuinely useful, not just a checkbox feature.
The $19.99/month for Gemini Advanced gets you Notebooks, all the core Gemini Advanced capabilities, Deep Research, and the NotebookLM connection. If you're using any two of those regularly, the math works out.
For students especially: if your institution's Workspace account gets access, this is probably the best document-centric AI setup available right now. The combination of Gemini's conversational ability and NotebookLM's source-grounding is more useful for academic work than anything else at this price point.
Disclosure: This article contains affiliate links to Gemini Advanced. If you subscribe through our links, we may earn a commission at no extra cost to you. This doesn't affect our coverage -- we only recommend products we've actually evaluated.
Top comments (0)