Verdict: Gemini Notebook's new Collections feature is the organization upgrade the tool has needed since its user base exploded past 30 million people. It works like Spotify playlists, not rigid folders — one notebook can live in multiple collections simultaneously, and nothing gets removed from your main dashboard. The feature started rolling out to all users on July 20, 2026, though the rollout is gradual and may take days or weeks to reach every account. The honest catch: Collections alone organize your notebooks but don't solve where the output of those notebooks goes. That second step — connecting your research to a durable knowledge system — is what separates people who use AI tools from people who actually build with them.
Last verified: July 30, 2026 · Collections rolling out since July 20, 2026 · Cloud code execution for AI Pro coming "over the coming weeks" · Pricing/limits may change — check Google's official plans page
What Are Gemini Notebook Collections?
Collections are a new organizational layer in Gemini Notebook — the source-grounded AI research tool formerly known as NotebookLM — that lets you group related notebooks the same way you group songs into playlists or photos into albums. You create a collection, give it a name (and optionally an emoji), select the notebooks you want inside it, and that's it. A single notebook can belong to as many collections as you want without being duplicated or removed from your main My Notebooks tab.
Google announced Collections on July 20, 2026, via the official Gemini Notebook account on X: "We heard you loud and clear: you want a way to organize your notebooks. So… Meet Collections! A brand new tab, Collections operate like photo albums or playlists: zero rigid folder structures with maximum flexibility."
The key design choice is that Collections do not behave like traditional file folders. A folder forces you to pick a single location — this file lives here and only here. Collections work differently: one research notebook about, say, customer onboarding can sit in a "Client Work" collection and a "Q3 Strategy" collection at the same time. No copies. No choosing sides. Nothing moves out of your main dashboard.
How to Set Up Your First Gemini Notebook Collection
Setting up a collection takes under a minute once the feature is available on your account. Here are the exact steps:
Open Gemini Notebook at notebooklm.google and look at the navigation tabs across the top. You should see My Notebooks, Shared with Me, and — if Collections has reached your account — the new Collections tab.
Click the Collections tab. If it's not there yet, you haven't done anything wrong. Google is rolling out gradually — check back in a few days.
Click "Create new collection" (or the equivalent button). Give it a name. Pick something outcome-focused — "Q3 Marketing Plan" tells you what to do; "Marketing stuff" tells you nothing.
Select the notebooks you want to add. You can choose as many as you want, and a notebook can belong to multiple collections.
Optionally add an emoji to the collection name. It sounds trivial, but when you have six collections on screen, a visual marker (📊, 🧪, 📝) helps you scan to the right one in seconds.
Hit Create. You're done. Your notebooks now appear grouped under that collection, and they also remain in your main My Notebooks tab.
Pro tip: Start with one collection for the project you're actively working on this week. Use it for a few days before building more. You'll quickly learn what naming patterns actually help you find things, rather than guessing your whole taxonomy upfront.
What Collections Can and Can't Do (July 2026)
Understanding the current limitations matters more than the advertising. Here's a clear-eyed breakdown:
| Feature | Status (July 2026) | Notes |
|---|---|---|
| Group notebooks into collections | ✅ Available | Rolling out gradually since July 20, 2026 |
| One notebook in multiple collections | ✅ Available | No duplication; no removal from main dashboard |
| Custom collection names + emojis | ✅ Available | Rename anytime |
| Share collections with others | ❌ Not yet | Personal-only; Google has shown interest in adding it |
| Nested collections (collections inside collections) | ❌ Not available | You get a flat list, not a hierarchy |
| Automatic sorting / AI tidying | ❌ Not available | You group things yourself — Gemini won't organize for you |
| Cross-collection search | ❓ Not confirmed | Not mentioned in the announcement or early coverage |
| Drag-and-drop adding notebooks | ❓ Not confirmed | Post launch— early users report selecting from a list |
The biggest gap is sharing. If your team collaborates on notebooks, Collections won't help distribute them — you'll still need to share individual notebooks manually. Google's official Workspace Updates post hints that sharing features are under consideration based on community feedback, but no timeline has been announced.
Why Google Renamed NotebookLM to Gemini Notebook
To understand why Collections arrived now, it helps to understand the bigger picture. On July 16, 2026, Google officially renamed NotebookLM to "Gemini Notebook." The tool was originally introduced as "Project Tailwind" at Google I/O 2023 and later launched as NotebookLM. The rename isn't cosmetic — it signals that the research tool is becoming part of the wider Gemini ecosystem, appearing inside the Gemini app and eventually inside Google Search's AI Mode.
The scale makes the organization problem obvious: Google reports more than 30 million people and over 600,000 organizations are now using the tool (blog.google, July 16, 2026). Business owners build onboarding materials with it. Students convert notes into audio summaries. Researchers analyze multi-document datasets. When a tool serves 30 million users and each can create dozens of notebooks, a flat alphabetical list stops being a feature and becomes a liability.
The Cloud Computer: What Else Shipped With the Rename
Collections are the visible upgrade, but Google also shipped a deeper capability alongside the rebrand: every notebook now gets a secure cloud computer that can write and run code natively (Google Blog, July 2026).
This means Gemini Notebook can now perform real data analysis directly grounded in your uploaded sources — cleaning inconsistent data, running calculations, and generating charts. The underlying model was upgraded to Gemini 3.5 with Antigravity (Google's coding tool), and the system includes more than 100 curated software skills (9to5Google, June 8, 2026).
Availability of the cloud computer:
| Tier | Code Execution Status |
|---|---|
| Google AI Ultra | ✅ Available now |
| Workspace Business (AI Ultra / Expanded Access) | ✅ Available now |
| Google AI Pro (web) | 🕒 Rolling out over coming weeks |
| Free tier | ❓ Not confirmed; check Google's plans page |
New output formats include downloadable files in PNG, SVG, PDF, DOCX, Markdown, CSV, JSON, XLSX, and PPTX — meaning your research can become a spreadsheet, a slide deck, or a chart without leaving the notebook.
How to Organize Your Research Sources Inside Each Notebook
Collections organize notebooks, but they don't fix a notebook stuffed with 20 random PDFs. The quality of answers in Gemini Notebook depends directly on the quality of your sources. Here's the framework that produces sharp, grounded answers:
1. Curate your sources deliberately
Gemini Notebook answers only from the sources you upload. If you dump 30 documents into one notebook, the AI has to search through noise to find relevant material. Instead:
- One notebook = one research question. A notebook called "Competitor analysis for product X" should contain the three competitor pitch decks, two pricing pages, and one industry report that answer that question — not your entire filing cabinet.
- Remove outdated sources. If a pricing page from January is no longer accurate, delete it or start a fresh notebook. Gemini Notebook includes the "Last verified" date in your output, so stale sources erode trust.
2. Use the Deep Research feature to discover sources
Instead of uploading files manually, you can have Gemini Notebook locate and add sources for you. In the Sources panel, click Add, then select Discover sources and type a research question. Gemini Notebook queries the web or your Google Drive, returns a multi-page report with cited sources, and lets you select which to import (Google support docs, July 2026).
Free accounts get 10 Deep Research runs per month; Google AI Pro users get 20 per day; Ultra tier allows 200 per day. Deep Research counts are tracked separately from your daily chat quota.
3. Keep notebooks to 10–15 sources for focused queries
The free plan allows 50 sources per notebook and 50 chat queries per day. Pro raises that to 300 sources and 500 daily chats. But more sources doesn't always mean better answers — for a specific question, 5-15 well-chosen sources outperform 50 loosely related ones (AI Weekly Guide, July 2026).
How to Connect Gemini Notebook Output to a Knowledge System (The Step Everyone Skips)
Here's the honest truth: Collections are a great first step, but organizing your notebooks doesn't solve the bigger problem — where does the output of those notebooks go?
Most people run AI tools like single-use apps: open one, get an answer, close it, forget it. The output disappears into a chat window you'll never scroll back through. A week later, you're doing the same work over again because you can't find what you already did. Building a knowledge system around Gemini Notebook involves three pieces — and many practitioners are already exploring how to connect research tools like this into broader AI agent workflows and agent OS architectures.
The three-layer system
| Layer | Tool | What it does | Why it matters |
|---|---|---|---|
| Input | Gemini Notebook + Collections | Research, analyze, and organize sources | Keeps your research structured and findable |
| Memory | A note-taking system (Notion, Obsidian, or similar) | Stores agent/notebook output so it persists | Work doesn't vanish into a chat scroll nobody revisits |
| Action | AI agent system (self-hosted or managed) | Runs agents that feed results into memory | Turns scattered one-shot answers into a process that compounds |
The principle is simple: tidy input, tidy system. Your notebooks are organized by Collections, your agent output is organized in a memory layer, and both talk to the same systems. If you're already running tools like a self-hosted Hermes Agent or experimenting with the latest Hermes Agent release, the same layering principle applies — the research tool feeds the agent, and the agent's output lands somewhere durable.
How to start (without overcomplicating it)
- Organize first. Create 2-3 Collections in Gemini Notebook around your active projects. Use them for a week. Naming tip: name collections after outcomes, not topics — "Q3 Competitor Battle Card" beats "Competitors."
- Pick a landing zone for output. When Gemini Notebook generates a summary, a chart, or an analysis, don't leave it in the chat. Export it (you now have PDF, DOCX, Markdown, and PPTX options) and drop it into a designated folder or note in your system — Notion, Obsidian, Google Drive, or wherever you actually return to.
- Connect your agents. If you use AI agents — Claude, Hermes, OpenAI — have them write their key outputs to the same memory location. Work that lands in a shared vault compounds; work that lives in a chat window dies.
- Review weekly. Spend 15 minutes every Friday checking: did anything useful get produced this week that I haven't routed to the right place? No system runs itself.
Gemini Notebook Plans and Limits (2026)
Pricing is a volatile fact — check Google's official plans page for the latest. Here's what was verified as of July 2026:
| Plan | Notebooks | Sources per Notebook | Daily Chats | Deep Research | Code Execution |
|---|---|---|---|---|---|
| Free (Standard) | 100 | 50 | 50 | 10/month | ❓ Not confirmed |
| Google AI Pro | 500 | 300 | 500 | 20/day | 🕒 Coming soon |
| Google AI Ultra | 500 | 500 | 2,500 | 100/day | ✅ Available now |
| Workspace Business | Included | Varies by tier | Varies | Varies | ✅ With AI Ultra access |
Sources: AI Weekly Guide (citing Google support docs), Google Blog, 9to5Google. Last verified: July 30, 2026.
Gemini Notebook is included in all Workspace plans at no additional cost, and your uploaded data is never used for model training (confirmed on the Google Workspace product page).
What This Means for You
If you're a small business owner: Use Collections to separate client research, vendor analysis, and internal strategy notebooks. One notebook with a client's documents can sit in a "Client A" collection and a "Q3 Decisions" collection simultaneously — you don't have to duplicate or choose where it "belongs." When Google eventually adds collection sharing, you'll already have the structure in place to hand off a curated set of research notebooks to a teammate.
If you're a researcher or student: The cloud computer feature is the high-value upgrade. If you're on Google AI Ultra (or have Workspace Business with AI Ultra access), you can now have Gemini Notebook write code to clean, analyze, and chart your data — all grounded in your own sources and exported as PDF, XLSX, or PPTX. Start with one notebook focused on one dataset and build from there.
If you're building with AI tools: The real leverage is the system, not the feature. Collections make it easier to find the right notebook. But the output of that notebook — the summary, the chart, the analysis — still needs to land somewhere that compounds. If you're running multi-agent pipelines or building an agent OS, your Gemini Notebook outputs should feed into the same memory layer your agents write to. Otherwise you're just collecting tools, not building a process.
FAQ
Q: How do I get Collections in Gemini Notebook?
A: Collections started rolling out to all users on July 20, 2026, but the rollout is gradual — Google does not enable it for every account on the same day. Open Gemini Notebook, check for a "Collections" tab next to "My Notebooks." If it's not there, wait a few days and refresh. There's no setting to toggle; it's a server-side rollout.
Q: Can one notebook belong to multiple collections?
A: Yes. That's the core design advantage over traditional folders. A notebook can be added to as many collections as you want. It stays in each collection and remains in your main My Notebooks tab — no duplication, no removal from any location.
Q: Can I share a collection with my team?
A: Not yet. Collections are currently personal — they organize your own workspace only. Google has acknowledged community requests for sharing and may add it in a future update, but no timeline has been announced. For now, continue sharing individual notebooks manually.
Q: Is Gemini Notebook the same as NotebookLM?
A: Yes. Google renamed NotebookLM to Gemini Notebook on July 16, 2026 (blog.google). It's the same product at the same URL (notebooklm.google). Existing notebooks, shared links, and bookmarks all continue to work through automatic redirects. No action is required.
Q: Can Gemini Notebook write and run code?
A: Yes, if you're on Google AI Ultra or a Workspace Business plan with AI Ultra or AI Expanded Access. Each notebook now has a "secure cloud computer" that can write and execute code for data analysis grounded in your sources. This feature is rolling out to Google AI Pro users on the web in the coming weeks. The underlying models were upgraded to Gemini 3.5 and Antigravity (9to5Google, June 2026).
Q: How much does Gemini Notebook cost?
A: Gemini Notebook has a free tier (100 notebooks, 50 sources each, 50 daily chats). Paid features are bundled into Google AI Pro and AI Ultra plans. Workspace subscribers get Gemini Notebook included in their plan at no extra cost. Check Google's official plans page for current pricing, as plans and limits change.
Q: What's the difference between Collections and folders?
A: A folder is a location — a file lives in one folder and that's its single home. A Collection is more like a label or playlist. A notebook can be in multiple Collections at once without being copied or moved from its original spot. You also can't nest Collections inside each other (no hierarchy), and there's no automatic sorting. The trade-off is simplicity: Collections are easier to use than a full file-management system, but less powerful for organizing large, structured libraries.
Discussion
0 comments