Without Leaving the Chat: Gemini Notebook Becomes a Connected App Inside Gemini
On 23 October 2026 Google announced that Gemini Notebook is now a Connected App inside the Gemini app: type @Gemini Notebook in a chat to create notebooks, add sources, list sources, or ask questions against your own material. This piece covers the four action groups, the default-on rule with its EEA exception, and how it differs from the 28 October Connected Apps platform preview.

No need to open another app: use your notebook right inside the chat
On 23 October 2026, Google announced on its Workspace Updates blog that Gemini Notebook is now also available as a Connected App inside the Gemini app. In plain language: you no longer have to leave the chat and open the notebook app separately to reach your research material. Inside a Gemini chat, typing @Gemini Notebook lets you create notebooks, add sources, save key notes, or ask questions and have them answered from your notebooks' contents, all in natural language. Gemini can also detect on its own when a prompt calls for the Notebook Connected App, so most of the time you do not even need the @ mention. Google stated the rollout pace plainly: available now for both Rapid Release and Scheduled Release domains, covering all Google Workspace customers, Workspace Individual subscribers, and users with personal Google accounts. Note what this is not: it is not a new product launch, but an upgrade in how you reach the notebook. The notebook already existed; what changed is where you can get at it.

Four things you can do in chat, each with Google's own example
The announcement spells out four groups of actions, each with a native example, and the examples are worth reading one by one because they reveal how Google imagines the tool being used. First, create, edit, or delete a notebook — the example is “Create a Gemini Notebook about solar farm market trends.” Second, add, update, or delete sources in a notebook — the example is “Add a copy of this chat to my notebook about Bio 101 course,” a telling one, because it means the conversation itself can become a source, so research no longer has to be tidied into a file before it can be saved. Third, list notebooks and their sources — the example is “Show me a list of all the sources in my notebook about US National Parks.” Fourth, answer questions using context from your notebook — the example is “What do my Gemini Notebooks say about butterfly migration?” Together, the four form a complete research loop: build the notebook, gather sources, check them at any time, and ask questions against them. For students and researchers, the most valuable combination is the second and the fourth: sources accumulate as you chat, and answers stay bounded by “my own material” instead of being fetched fresh from the whole web each time.
First, a distinction: in the side panel versus usable inside the chat
One confusion needs clearing up first. According to 9to5Google's same-day reporting, Gemini notebooks had already appeared in the Gemini app's side panel; what is new is the deeper Connected App integration. What is the difference? The side panel is a place you go to look at notebooks; the in-chat integration is a capability you use where you are already talking. The first is an entrance, the second is a power: previously you shuttled between the side panel and the conversation, with sources and questions living in two interfaces; now the question can travel with its source context, and answers are scoped to the notebooks you point at. That difference matters for real work like writing a paper, revising for a course, or preparing a report: you are not gaining another place to view notes, but losing one manual step of “copying sources into the question.” Honesty requires two caveats: the announcement says nothing about pricing or usage limits, and nothing about mobile availability — 9to5Google, citing the help page, describes it as web-only for now. By the announcement's own terms, only the “available now” rollout pace and the covered audiences are promised; anything beyond that should be checked against the help page rather than assumed.
On by default — except in Europe, where it is off: admins and individuals each own a switch
The rollout's switch policy has two layers, and individuals and admins each own one. For individual users, the feature is ON by default globally, with a single exception: in the European Economic Area it is OFF by default, and users there must turn it on themselves under Settings and then Connected Apps in the Gemini app. For Workspace admins, access is controlled by the Gemini and Gemini Notebook access settings together: the Connected App is available to users in a group or organizational unit where both are set to On. In other words, in a company or school setting, an individual user still needs the admin to enable both before the feature works; conversely, if an admin switches either one off, the whole integration stops working. This design matters because it determines real availability inside teams: when the announcement says it covers “all Workspace customers,” that is about eligibility, not “everyone can open it tomorrow.” If you cannot find @Gemini Notebook on a company or school account, asking the admin whether both settings are on will help more than reinstalling the app.
This is not the same story as the 28 October “Connected Apps platform”
Google had two “connected” stories in late October, and they must be kept apart or they will read as one article. This site covered the Google Connected Apps platform preview in cycle128, dated 28 October: a platform for developers, with cloud MCP and Android AppFunctions as the two integration routes, a Connected Console, and more than 50 apps, aimed at letting third-party apps plug into Gemini and Search. This article's 23 October news is one specific, end-user integration under that same idea: Gemini Notebook, a single app, becoming a Connected App that can be used directly inside Gemini chats. One is building the road; the other is the first vehicle driving on it. Seeing the relationship explains Google's sequencing: bring its own notebook in first as the model case, prove the experience of calling on your own material inside a conversation, and then extend the pattern to more third parties. For ordinary users, the platform news is the distant view and this article is the thing to try tomorrow; for developers, the order reverses, and the platform news is the signal that decides whether to follow.
Three kinds of users, three ways to use it: students, researchers, librarians
The same feature becomes three different tools in different hands, so it helps to match yourself to one before deciding whether to try it. First, students: with a course notebook holding lecture notes, paper excerpts, and their own class notes, revision no longer means reopening files — ask in chat “what are the three key experiments on photosynthesis in this notebook” and the answer's context is your own material, so the revision scope cannot drift. Second, researchers working on a topic: sources are messy — papers, interview notes, public data documentation — and the old workflow was to paste source summaries into the chat before asking; now you can ask “which sources in my notebook disagree about the sample size, and what does each say,” letting the conversation search the notebook's source list — while still checking each surfaced source yourself, since the tool saves the finding step, not the verification step. Third, the team member who maintains a shared notebook, such as a product research library: when a new hire joins, instead of reading the whole library first, they can ask in chat “what conclusions about competitor pricing does this notebook hold” to build the big picture before reading the originals closely. All three share one precondition — a notebook that already exists. If you hold no finished notebook at all, the feature means little to you; build one notebook with more than twenty sources first, and only then judge the in-chat experience, or you are simply using a more expensive chat box.
What it cannot do: four boundaries, stated up front
Stating the boundaries up front beats disappointment later. First, it does not judge whether your sources are right. Material in a notebook can be outdated, one-sided, or something you saved hastily without reading; when Gemini answers from it, it will not automatically fact-check for you, and the answer's credibility ceiling is your sources' credibility. Second, the announcement does not say whether answers will cite their sources, nor does it promise answers are strictly confined to notebook contents, so in practice you should ask “which source does this sentence come from” yourself, and never paste a chat answer into a paper or report as if it were a citation. Third, it does not change the notebook's own management rules: adding, removing, sharing, and storing sources still follow the notebook app's own logic; the chat is an additional operating entrance, not a replacement permission system. Fourth, the announcement says nothing about price or usage caps; for individual users there is currently no statement of an extra charge, but equally no written promise of “free forever” — by the announcement's own terms it is an integration inside the existing Gemini app, not a separately sold feature, and cost questions beyond that belong to the help page. Keep these four in mind when trying @Gemini Notebook, and you will not expect more from one integration than it can deliver.
Same tool, new name: the July renaming behind the story
One more piece of background is worth stating, so readers are not confused by the name. According to multiple reports, Gemini Notebook was previously NotebookLM, and Google renamed it Gemini Notebook on 16 July 2026, repositioning it from a standalone research tool into a research assistant spanning the Gemini app and Search. The rename did not rebuild the core capabilities — uploading sources, generating summaries, asking questions against sources — and this 23 October update completes “calling it inside Gemini chats” under the new name. The timeline helps twice: first, tutorials and discussions about NotebookLM from before July describe essentially the same tool, so do not skip them just because the name does not match; second, the July rename settled the identity question (“it belongs to the Gemini family”), while this October update settles the experience question (“can it be used inside Gemini chats”), and both moves point the same way — bringing the notebook closer to the conversation. For existing users, there is nothing new to relearn, just one more thing to try in chat: @Gemini Notebook. For new users, now is actually the easiest moment to start, because the name, the entrance, and the calling method have been unified into one set.
Three steps to start: from a first notebook to a first in-chat call
If you want to try it now, three steps waste the least time. Step one: build a decent notebook first, with as specific a topic as possible — “Biology 101 midterm revision” rather than “study material” — and put more than ten sources in it: lecture notes, paper excerpts, your own notes all count. With too few sources, the in-chat experience is hard to judge, because the answers will look no different from asking directly. Step two: call it by name in a Gemini chat with @Gemini Notebook, and first ask a question you already know the answer to, such as “what is this notebook's definition of a certain concept,” then check whether the answer genuinely comes from your material and whether its wording drifted. If the answer clearly comes from general knowledge rather than your notebook, either the call did not take effect or the notebook is too thin — add sources before trying again. Step three: build the habit of turning conversations into sources: after a valuable conversation, ask Gemini to save a copy of it into the matching notebook, so that conversation becomes quotable material for the next question. The step most often skipped is the check in step two; many people try it once, think “that answered well,” and start trusting it, but “answered well” and “answered from my material” are two different things, and research work must tell them apart. Once all three steps work, organize notebooks by course, topic, or project instead of stuffing everything into one catch-all — the more focused the notebook, the more accurate the in-chat answers.
Judgment: it attacks the old problem of “sources scattered everywhere”
Why does this integration deserve its own article? Because it targets not “a smarter model” but the most grinding part of research work: sources scattered in one place, questions and answers in another. The real workflow for many people looks like this: sources sit in the notebook, questions have to be re-described from scratch in the chat box, and answers then have to be checked against sources by hand. The Connected App compresses all three steps into one conversation: the sources are already there, the question is asked directly, and the answer's context is the notebook itself. The change is small, but it belongs to the “saves ten minutes a day” class, and those add up. Its limits are equally clear: it does not judge sources for you, it does not design the research for you, and it does not promise answers from authoritative sources beyond the ones you specified; it only makes “asking questions against your own material” cost one less jump. The advice for Gemini Notebook users is one sentence: take an existing notebook, ask it three questions you already know the answers to in chat, confirm it is really answering from your material, and only then fold it into your daily research flow. For people who have never used the notebook, the news means “start with one small notebook” — not “move all your material in at once.” And once that flow runs smoothly, consider bringing a second and third notebook in as well, so the chat becomes a single entrance to your personal material rather than a one-off tool for a single course or project.
Comments (0)
Sources
Related articles

On 28 October 2026 OpenAI's API changelog announced free monthly API credits for eligible ChatGPT Pro and Business subscribers: $100, $200, or $500 by Pro tier, and seat-pooled credits for Business capped at $500. This piece unpacks the allowance table, the three claim requirements, the Codex exclusion, four traps, and why this is a different ledger from the chat-limit doubling.

On 28 October 2026 GitHub made repository plugin defaults and enterprise plugin controls generally available: repos can pre-install default plugins, enterprises can require, block, and restrict to approved marketplaces, and enterprise policy wins over repo and personal choices. This piece checks scope and precedence item by item, with a rollout order for three roles.

GitHub's 28 October 2026 changelog says Copilot can now generate and edit images in chat with natural language, previewing inline and saving to the project. This piece works through billing, watermarking, and checking: what it removes is the switch, not design itself.