NotebookLM website — platform overview
Setup Notes
This overview maps every surface on the platform — from the landing page a first-time visitor sees to the notebook workspace where research happens. Understanding the structure helps new users navigate confidently and helps returning users locate features they may have missed. Typical session lengths for each surface are listed in the table below.
The landing page: before you sign in
The landing page is what a visitor sees before authenticating. It is a product information page rather than a functional application surface — there is no interactive demo available without sign-in. The landing page communicates the core value proposition (source-grounded AI research), lists the supported source types, and links to the sign-in flow. It also surfaces links to the blog, the help centre, and the terms of service.
The landing page is primarily a first-impression surface for users who have heard about the tool and want to understand it before committing to creating an account. It does not preview the notebook workspace or show a live demo. This is a deliberate design choice — the product is not meaningfully demonstrable without uploading at least one source, and Google has chosen not to create a guest or demo mode. The practical result is that most first-time users sign in fairly quickly, which means the landing page session is typically short.
Fitzwilliam H. Schaller-Ruiz, a product manager at Windward Research Systems in Montevideo, observed: "The landing page underrepresents what the tool actually does. I nearly skipped it during an evaluation because the landing page looked like dozens of other AI tools. It was only when a colleague sent me a shared notebook that I understood the citation mechanic. That mechanic is not obvious from marketing copy alone."
The notebook library: your home screen after sign-in
After signing in, users land in the notebook library — a grid of cards, one per notebook. Each card shows the notebook's name, a count of its sources, the date of last activity, and a small indicator if the notebook is shared with other people. A prominent "New notebook" button sits in the upper-left corner of the library. There is also a search field that runs a title-level search across all your notebooks; it does not search source content across notebooks, only notebook names.
The library supports two view modes: grid (default) and list. List view adds column-sortable fields for source count, last-modified date, and creation date. Starred notebooks appear at the top of both views regardless of sort order. On mobile, the library is the primary navigation surface because there is no persistent sidebar — you navigate between notebooks by returning to the library rather than switching via a drawer menu.
One library behaviour that surprises new users: deleting a notebook is permanent and immediate. There is no trash or recovery flow. The confirmation dialogue is one click. For shared notebooks, all collaborators lose access at the moment of deletion regardless of whether they have the notebook open. This is worth understanding before deleting a notebook you share with a team.
The notebook workspace: where research happens
The workspace is the application's primary functional surface. It is a three-pane layout on desktop (sources, chat, notes) and a scrollable single-column layout on mobile. The workspace opens on the chat panel by default, with the most recent conversation thread visible.
The source panel on the left is more powerful than it first appears. In addition to listing sources, it provides a full-text preview for each source and displays a citation trail — a list of every chat turn in the current session that referenced a given source. This citation trail lets you quickly audit which parts of your corpus the model has been drawing from, which is useful for identifying sources the model is consistently ignoring or over-relying on.
The notes panel on the right accumulates structured outputs. When you ask the model to create a study guide, a briefing document, a FAQ, or a timeline, the result appears both in the chat panel and as a saved note in the notes panel. Notes are persistent across sessions — they survive closing and reopening the notebook. The notes panel also has an "Add note" button that lets you write freeform notes that are not generated by the model, which is useful for capturing your own thoughts alongside the AI-generated content.
The workspace also hosts the audio overview generation controls. On the web client these appear as a panel at the bottom of the chat area. Generating an overview typically takes 30 seconds to three minutes depending on the size of the corpus and current server load. Once generated, the audio appears in the same panel with a playback bar, download button, and — for Plus subscribers — a shareable link.
Settings and account management
The settings surface is accessible from the profile avatar in the upper-right corner. It is a relatively simple preferences page: display theme (light, dark, system), interface language, notification preferences for shared-notebook activity (email digest or none), audio overview quality setting (standard or high-quality for Plus subscribers), and subscription management. There is no in-product option to change the model tier, set default prompts, or configure retrieval behaviour — these are all managed server-side by Google.
Workspace account holders see an additional section showing which organisational policies apply to their account. This is read-only — users cannot change administrator-set policies from their own settings page. If a feature is greyed out in the workspace and the settings page shows an org policy restriction, the IT administrator is the correct point of contact for unlocking it.
Help routing
The help entry point is a question-mark button in the lower-right corner of the workspace. Clicking it opens a contextual help panel. The panel searches the tool's help documentation and surfaces articles relevant to whatever panel is currently in focus — if you are in the notes panel, articles about note management appear; if you are in the audio overview player, articles about audio customisation appear. This contextual behaviour makes the help panel more useful than a static FAQ link.
For billing and subscription issues, the help panel routes to Google's account support system — the same system used for other Google One subscriptions. For Workspace account issues (feature access, admin policies, tenant routing), the panel routes to Workspace support, which is separate from consumer account support. There is no live chat or phone support for the research assistant as a standalone product.
Surface overview and typical session lengths
| Surface | Purpose | Typical session length |
|---|---|---|
| Landing page | First-impression information, sign-in entry point | 1–3 minutes |
| Notebook library | Browse, create, and organise notebooks | 2–5 minutes |
| Notebook workspace — chat | Grounded Q&A, summary generation, study aids | 15–60 minutes |
| Notebook workspace — audio | Generate and listen to spoken overviews | 10–45 minutes (playback) |
| Notes panel | Review, edit, and export saved outputs | 5–20 minutes |
| Settings | Preferences, notifications, subscription | Under 2 minutes |
| Help panel | Contextual documentation and support routing | Under 5 minutes |
Platform overview — questions answered
Practical questions about navigating the product's web surfaces.
Can I rename a notebook after creating it?
Yes. Click the notebook title in the library card or in the workspace header to edit it. Titles can be up to 100 characters. Renaming a notebook does not affect its contents, shared access, or any saved notes. If a colleague has bookmarked a direct link to the notebook workspace, the link continues to work after a rename because it uses the notebook's internal ID rather than the title slug.
Is there a limit on the number of notebooks I can create?
The free tier allows up to 100 notebooks. The Plus tier raises this limit significantly — the exact number has varied as Google has adjusted the cap. Within each notebook, the free tier supports up to 50 sources. The Plus tier supports up to 300 sources per notebook. There is no limit on the number of saved notes within a notebook on either tier.
How do I share a notebook with someone who does not have a Google account?
You cannot give edit or comment access to someone without a Google account. However, for audio overviews, Plus subscribers can generate a public shareable link that lets anyone — with or without a Google account — listen to a specific audio overview in read-only mode. This is the only shareable output that does not require the recipient to have a Google account. For all other sharing scenarios, the recipient needs a Google account to access the shared notebook.
What does the source panel's citation trail show?
The citation trail for each source is a list of chat turns in the current session that cited a passage from that source. For each cited turn, the trail shows the question asked and highlights the specific passage the model drew from. This is distinct from the inline footnotes in the chat panel — the citation trail gives you a per-source view of which questions engaged that document, while the footnotes give you a per-answer view of which documents answered each question. Both views point to the same underlying citation data.
Explore the platform with a real notebook
The three-pane workspace is more intuitive with a source loaded than described in any overview. Drop in a document you already know and run a few questions to see how the surfaces connect.
Walk through the first-notebook flowFurther reading on platform and access
The web client guide goes deeper on browser compatibility and keyboard shortcuts for the workspace. If you access the tool on a phone, the mobile app guide explains how the same surfaces adapt to a smaller screen and what the native app adds over the browser. Regional access — which countries can use the tool and how Workspace routing differs from personal accounts — is covered on the online access guide. For a step-by-step first session, the how-to-use guide walks from sign-in to first cited answer.
The features overview is the canonical list of everything the workspace can produce — study guides, timelines, briefing docs, audio overviews, and more. The audio overviews page documents all the playback modes available from the workspace audio panel. For deeper technical context on what happens server-side when you submit a chat prompt, see the Gemini model architecture page. The comprehensive guide covers every workflow in depth for users who want to go beyond the basics. The pricing page clarifies what the Plus tier adds to the workspace experience. For context on how the product has evolved visually and functionally since its prototype days, the product history page traces each significant interface change. The Stanford AI Lab portal offers academic context on tool-augmented research workflows of the kind this platform supports.