How to Organize Gemini Notebook (formerly NotebookLM) With Folders

You can organize Gemini Notebook (formerly NotebookLM) in four moves: a naming rule, a few collections, three kinds of tags and a short weekly review. Why bother? Picture this. You open Gemini Notebook (the new name for NotebookLM) and see 67 notebooks. Half are called “Untitled notebook” or “Research March 2026.” You can’t recall what most of them hold, so you scroll, give up and make notebook number 68. That’s the normal end state of heavy use, and it arrives fast. In fact, when we analyzed 776 Kortex MCP waitlist answers in September and October 2026, we found that 327 people (42%) keep more than 50 notebooks. This guide gives you the system that keeps a library that size findable.

Key Takeaways

• Gemini Notebook’s native Collections, rolling out since July 2026, group notebooks like playlists but add no tags or nested folders.
• A three-part name, [TYPE] Topic: Subtopic, fixes most navigation pain with no extra tools.
• Three tag tiers (status, type, priority) turn a long scroll into a few clicks.
• Kortex, the extension we built, adds folders with subfolders, colored tags and rules that file notebooks for you.

Why do most Gemini Notebook workspaces become unmanageable?

Most workspaces get messy because Gemini Notebook makes notebooks easy to create and gives you little to file them with. The free plan allows 100 notebooks, Google AI Plus allows 200 and AI Pro allows 500 (Gemini Notebook Help, 2026). That is, in fact, a lot of room for clutter. Meanwhile, each notebook is a closed box: you can’t search one notebook from inside another, and there are no tags. Native Collections, rolling out since July 2026, group notebooks “like playlists on Spotify,” and one notebook can sit in several of them (Android Authority, 2026). However, Collections can’t be shared, and they don’t nest. As a result, the default view is still a long list sorted by date. Five notebooks is fine. Fifty means every session starts with a hunt.

The demand is real, too, and it comes from heavy users in particular. Of the 776 people we surveyed, 121 asked for help to tidy and organize their library, and half of them already pay for Kortex Pro. One wrote that they search “using Ctrl+F, but notebook names change unpredictably, forcing me to review hundreds of them one by one.”

The cause is simple. Specifically, Gemini Notebook was built as a research tool, not a file manager, so order has to come from you. Consequently, the good news is that the system that works is small and takes about 10 minutes a week once it’s set up.

What is the most effective naming convention for Gemini Notebooks?

The most effective convention is [TYPE] Topic: Subtopic, typed the moment you create a notebook. It’s the single biggest change you can make, and, importantly, it costs nothing. A good name does three jobs. It tells you the kind of content at a glance, it names the topic, and it lets you find the notebook by typing part of the name. Consider what happens in a sorted list: every [CLIENT] notebook lands next to the others, and every [BOOK] notebook does the same. In other words, the brackets act like folders without any folder feature. Names matter more than you might expect, however. A Google Research review of personal search (Bendersky et al., arXiv, 2024) notes that some people skip filing altogether and rely on search to find things again. If that’s you, clear names are what your search will hit. On the other hand, if you’re a filer, they’re what you’ll scan. Either way, a name you can read beats a name you have to open.

The format that works across most use cases:

[TYPE] Topic: Subtopic

Type prefixes to use:

• [BOOK] for a book you’re reading and annotating.
• [PROJECT] for an active work or research project.
• [COURSE] for a course, lecture series or tutorial.
• [CLIENT] for client research or briefings.
• [PERSONAL] for side projects and learning.
• [RESEARCH] for papers and primary sources.

Examples:

• [BOOK] The Innovator's Dilemma: Clayton Christensen.
• [PROJECT] Product launch: Q2 2026 strategy.
• [COURSE] Stanford ML: weeks 3 to 6.
• [CLIENT] Acme Corp: due diligence.
• [RESEARCH] Climate policy: IPCC Sixth Assessment.

As a result, the subtopic after the colon lets you keep several notebooks on one topic without a clash.

The most important rule: name the notebook when you create it, not later. “Untitled notebook” means you must open it to know what’s inside. In contrast, a clear name means you know before you click. It helps AI assistants too. For example, our guide to using your notebooks as long-term memory for Claude or ChatGPT starts by mapping a library like this one through Kortex MCP, which is a private beta for invited Kortex Pro members.

How do you build a tag system that scales?

Build it in three tiers: exactly one status tag, one or two type tags, and a priority tag only when it applies. Tags are labels that describe a notebook’s state, while names describe its content. That split matters. For instance, a notebook named [CLIENT] Acme Corp: Q1 review tells you the topic, but what you need right now is whether it’s active, finished or waiting on you. Gemini Notebook has no notebook tags of its own, so this layer comes from Kortex, which adds colored tags you can apply to many notebooks at once, rename or delete. The payoff shows up at scale, specifically once you pass a few dozen notebooks. With 60 notebooks and no tags, finding “every active client notebook with a deadline” means scrolling and guessing. With the three tiers below, it’s three clicks on a filter. Notably, the status tier does most of the work: it tells you whether to open a notebook or skip it, which is the choice you make dozens of times a week.

Tier 1: Status tags (always exactly one)

Status tags answer “what stage is this in?”

• #active: you’re working in it this week or month.
• #reference: the project is done, but you’ll look things up in it.
• #archive: done, and you’re unlikely to need it again.

Tier 2: Type tags (one or two)

Type tags answer “what is this for?”

• #research: papers, literature reviews, primary sources.
• #client: client work.
• #personal: personal projects and learning.
• #course: courses, tutorials, certifications.
• #writing: articles, newsletters, book drafts.

Tier 3: Priority tags (only when relevant)

Priority tags answer “is anything urgent here?”

• #urgent: a deadline is close.
• #review: waiting for your review.
• #shared: shared with a colleague or client.
• #stale: you haven’t touched it in a month (you add this during the weekly review).

Example combinations:

• [CLIENT] Acme Corp: due diligence tagged #active #client #urgent.
• [RESEARCH] Climate policy: IPCC tagged #reference #research.
• [COURSE] Stanford ML: weeks 3 to 6 tagged #archive #course.

How do you group notebooks into folders or collections?

Use five to eight stable collections and let tags handle the rest. Collections are named groups of notebooks. Google’s native version works like a playlist, while Kortex’s version works like folders, with subfolders, colors and pinning. You can also make smart collections, which are folders that fill themselves: you set tag rules once, and any notebook that matches all of them joins on its own. However, keep the count low. The same Google Research review found that among “spring cleaners,” people who file in bursts, over half of the folders they made turned into failed folders they rarely used again (Bendersky et al., arXiv, 2024). Therefore, a folder per project is a trap. It rebuilds the flat-list problem one level down. Instead, pick broad, lasting groups and use tags to tell notebooks apart inside them. Five to eight collections suit most libraries, and three is a fine start.

A structure that fits most people:

• Active Projects: notebooks you’re working in now.
• Client Work: every client notebook, whatever its status.
• Reference Library: finished research you’ll look up again.
• Courses: learning material from classes and tutorials.
• Archive: finished projects waiting to be exported and cleared.

Here is how the two kinds of collections compare.

	Native Collections	Kortex collections
Subfolders	No	Yes
Colors and pinning	No	Yes.
Fill themselves from tag rules	No	Yes (smart collections)
Notebook tags	No	Yes.
Cost	Included	Free plan has 5 folders; Pro is unlimited.

If you run 50 or more active notebooks, the Kortex vs Gemini Notebook comparison shows how search across collections and tags compares with the default list.

What does an effective weekly review look like?

A 10-minute weekly review keeps the system honest. Without it, tags drift out of date as projects change; therefore, put it on your calendar. Run it every Monday or Friday, whichever suits you:

1. Scan your #active notebooks. Anything you haven’t opened in two weeks moves to #reference. If you still mean to work on it, the project is paused, not active.
2. Check your #reference notebooks. Anything untouched for three months moves to #archive. Reference is for things you look up. Archive is for things you keep but won’t use.
3. Tag what’s gone quiet. Sort by date added and give #stale to anything you haven’t touched in a month. Then decide: keep, move to reference or archive.
4. Process your #archive. Export any archived notebook you might want outside Gemini Notebook, then keep it or delete it. Exporting first means, in other words, that nothing is lost by accident.

Similarly to tidying a desk, the review gets faster the more often you do it. A clean library this week makes next week’s pass shorter.

Can you organize Gemini Notebook automatically?

Yes, partly. Kortex Pro includes automation rules with 8 triggers, 11 actions and 9 ready-made rules, so filing can happen without you. We built them so that one event, such as adding a tag, sets off the next step. In particular, three of them fit this system. Smart Sort is a rule that moves a notebook into a collection when it gets a tag: tag it #client and it lands in Client Work. The Escalator adds more tags when one tag appears; for example, tagging #paper can also add #to-summarize and #cite. A third pattern pairs a tag with an artifact: tag #to-summarize and a briefing doc and flashcards start generating. That said, there are limits. Kortex has no “not opened in 30 days” trigger, so spotting stale notebooks stays part of the weekly review. Additionally, rules never delete a notebook on their own. Automations are a Pro feature; the free plan gives you 5 folders and the manual system above.

For longer chains, the complete automation workflows guide has 10 patterns with exact trigger-and-action setups for researchers, writers and knowledge workers. If you also keep a prompt library, the 30 Gemini Notebook prompt templates use a category structure that maps onto the collections above.

What are the most common organization mistakes?

Most failures trace back to two habits: too many collections and skipped names. Keep your collections under eight and name every notebook the moment you make it. Those two rules prevent most long-term mess.

Creating collections for every project. Collections should be lasting categories. If you make one called “Acme Corp Q1 Review,” you’ll have 40 collections within months. Use tags for project labels instead; likewise, use subfolders sparingly.

Skipping the naming step. The name “Research” is useless at 50 notebooks, and “Untitled notebook” is worse. Naming takes 15 seconds, and skipping it costs you every time you open the list. And if notebooks seem to vanish, check why Gemini notebooks go missing before you rebuild anything.

Using type tags without status tags. Knowing a notebook is #research helps less than knowing it’s #archive #research. Consequently, the status tag is the one you can’t skip.

Never deleting anything. Notebooks don’t cost you storage in any real sense. Nevertheless, a library that only grows gets slower to scan, and the free plan stops at 100 notebooks. A yearly clear-out of old archives keeps it usable.

How do you clean up a messy Gemini Notebook workspace?

Start from where you are: rename, tag, file, then export and delete, in that order. If you have 60 messy notebooks rather than none, the four phases below take about an hour in total. Do them in one sitting if you can, because momentum matters more than polish here. Notably, the order is deliberate. Renaming first means every later step works on names you can read. Tagging second means filing becomes a filter, not a judgment call per notebook. Exporting before deleting means nothing is lost by mistake, and Kortex can save a whole notebook’s sources as PDF, Markdown, text or a ZIP with one button. Also note that sources still count against caps you can’t lift by tidying: the free plan holds 50 sources per notebook, each up to 500,000 words or 200 MB (Gemini Notebook Help, 2026). Thus a cleanup is also a good moment to split overstuffed notebooks.

Phase 1: Rename (20 minutes). Sort notebooks oldest first and rename each one to the [TYPE] Topic: Subtopic format. If you can’t recall a notebook’s content, open it, read the source titles and name it.

Phase 2: Tag (20 minutes). Give every notebook at least a Tier 1 status tag. Most will be #archive, some #reference and a few #active. Don’t overthink it.

Phase 3: File into collections (10 minutes). Assign each notebook to a collection. Start with three: Active, Reference, Archive. You can add more once the system proves itself.

Phase 4: Export and delete archives (10 minutes). Export the #archive notebooks you want to keep, then delete the ones you’ll never open again.

Similarly, usage limits won’t block the cleanup. Since September 2, 2026, Gemini Notebook’s usage refreshes every five hours up to a weekly cap (Gemini Notebook Help, 2026), and the 100-notebook cap stayed the same when Google switched to compute-based usage limits. After the cleanup, the getting started guide shows where the automation rules live.

Frequently asked questions

Does Gemini Notebook have folders?

Not real folders. Since July 2026, Gemini Notebook has been rolling out Collections, which group notebooks like playlists, and one notebook can belong to several. At launch, Collections couldn’t be shared, and there are still no nested folders or notebook tags. Kortex adds collections that work as folders, with subfolders, plus color-coded tags.

Does this organization system work without Kortex?

Partly. The naming convention works in plain Gemini Notebook, and native Collections give you basic groups. What you don’t get without Kortex is tags, subfolders, smart collections that fill themselves from tag rules, automation rules and one-click export of a notebook before you clear it out.

How many notebooks before organization breaks down?

There’s no fixed number, but the free plan caps you at 100 notebooks and each one at 50 sources, so a busy library hits the wall sooner than you’d think. In our survey of 776 Kortex MCP waitlist members, 42% said they keep more than 50 notebooks. Past that size, a flat list with no tags gets slow to scan.

Can I search inside notebooks from the organization view?

Kortex’s search box looks through notebooks, sources, notes, artifacts, collections and tags, which is usually enough to find the right notebook. To ask a question across the text of several notebooks at once, Kortex MCP’s ask_library tool queries up to three notebooks; it is a private beta for invited Kortex Pro members.

How often should I do the weekly review?

Weekly is ideal. Monthly works if you’re a lighter user. The longer you go between reviews, the more time the review takes, so a short weekly pass usually beats a long monthly one.

Does Kortex delete notebooks automatically?

No. Kortex rules file, tag and generate things for you, and deleting a notebook stays a step you take yourself. The one rule action that removes anything is Delete sources, and it only runs if you add it to a rule on purpose.

---

Editorial note: fact-checked against the official help pages and docs linked above on October 2, 2026. Spotted something out of date? Contact us and we will fix it, or read more about Kortex.