On April 8, Google announced the launch of Notebooks within Gemini, directly integrating NotebookLM into Gemini's primary workflow.
On the surface, this is a product update. Behind it lies an important judgment from Google: the future of AI assistant competition is not about which model answers questions more accurately, but which can better manage users' personal knowledge systems.
How this feature actually works
The Gemini sidebar now has a new Notebooks entry. You can put the following into it:
- PDF documents
- Files from Google Drive
- Web links
- YouTube videos
- Directly copied and pasted text
Once a notebook is created, when Gemini answers your questions, it will simultaneously use the content you put in the notebook and web search results to generate a comprehensive answer — rather than relying solely on the model's own knowledge.
A more critical point is that these notebooks are automatically synced bidirectionally between Gemini and NotebookLM. Notebooks you create in Gemini appear directly in NotebookLM. Source files you add in NotebookLM are also usable in Gemini.
Previously, these two products were completely separate. Now they are connected.
Deep Research is also integrated
Google has simultaneously done another thing: the results of Deep Research can now be saved directly into a notebook.
Deep Research is Gemini's in-depth research feature — give it a question, and it automatically searches dozens of sources, outputting a structured report. Previously, this report was gone after use. Now it can be saved, along with all the sources it cited, into a notebook for continued use with NotebookLM for audio summaries, video overviews, and follow-up questions.
With this chain connected, a standard research workflow could be:
- Run a background research session with Deep Research
- Save the report and sources together into a notebook
- Continue the conversation with follow-up questions in the notebook
- Use NotebookLM to generate an audio Overview for quick digestion
If this chain works smoothly, the efficiency improvement for research-type work could be significant.
How it differs from OpenAI/Anthropic Projects
All three companies are moving toward persistent context management, but the logic is not exactly the same.
OpenAI's Projects and Claude's Projects are more focused on remembering your work habits and preferences — a cross-session long-term memory system.
Google's Notebooks are more focused on managing your knowledge sources — whatever content you put in, Gemini uses that content to answer you, with clear boundaries.
This reflects two different use cases:
- Doing research, writing, and analysis that requires managing a large amount of raw material → Notebooks is more suitable
- Using AI for daily work assistance, needing it to remember your context and preferences → Projects-type products are more suitable
It's not surprising that Google chose the research scenario to start with — NotebookLM was originally built from this scenario, has been running for two years, and has accumulated many specific features: audio summaries, video overviews, FAQ generation. These are differentiating capabilities.
A detail worth noting
There is a feature called "Use notebook memory" that allows Gemini to use the content from your notebook to enhance answers in all conversations — not just within the context of a specific notebook, but globally.
This is a significant change: previously, files imported into AI disappeared when the conversation ended. Now you can build a persistent private knowledge base that influences all of Gemini's answers.
You can set custom instructions for each notebook, specifying the tone, format, and focus areas for Gemini's answers. This level of customization is a real efficiency improvement for long-term projects and professional research.
Access restrictions and rollout pace
Currently, Notebooks is only available to Google AI Plus, Pro, and Ultra subscribers, initially on the web version. It will later expand to:
- Mobile devices
- More European countries
- Free users
NotebookLM is currently running on the Gemini 3 model.
The trajectory of NotebookLM from an experimental product to being integrated into Gemini's main product line indicates that Google believes this product direction is correct. What is worth watching next is whether this Notebooks system can get more ordinary users to actively use Gemini, rather than just occasionally asking about the weather or doing translations.
Sources: Google introduces Notebooks in Gemini, a project management tool synced with NotebookLM (Google Blog); Gemini app rolling out 'notebooks' to organize chats & files, CocoLoop, integrates with NotebookLM (9to5Google); Google Integrates NotebookLM Into Gemini App For All Users (Dataconomy); NotebookLM adds Deep Research and support for more source types (Google Labs Blog)