If you bookmarked notebooklm.google.com at some point and have not opened it in a while, the bookmark still works, it just redirects you somewhere else now. Google renamed NotebookLM to Gemini Notebook on July 16, 2026, moved it to notebook.google, and folded in a handful of features that go well beyond a cosmetic name swap. The tool that started as a research assistant for summarizing your own documents can now write and execute code inside a notebook, pull outside web sources into your research automatically, and export charts, spreadsheets, and slide decks. None of that shows up unless you actually test it against the free plan's real limits, which is where the gap between "major upgrade" and "still capped" becomes obvious.
The rebrand itself is simple, the reason behind it is not
Functionally, nothing about your existing notebooks changed. Google has said existing users do not need to do anything, old notebooks still work, and the old URL redirects automatically. The actual move is strategic: Gemini Notebook is Google folding yet another standalone AI product into the Gemini brand, the same pattern it has used with several other tools this year. The product itself traces back to Project Tailwind, first shown at Google I/O in 2023, and by the time of the rename it had grown to roughly 30 million users and more than 600,000 organizations using it, according to Google's own figures.
The practical effect for a regular user is mostly about discovery. Gemini Notebook now shows up as an option inside the main Gemini app itself, instead of living as a separate, easy-to-forget tab you had to remember existed.
The cloud computer is the real upgrade
The feature worth paying attention to is not the name, it is what Google calls a secure cloud computer attached to every notebook. In practice, this means Gemini Notebook can now write and run code against your uploaded sources rather than just summarizing them in plain text. Google says this is backed by more than 100 curated software skills and has posted win-rate numbers against its own prior version: a 69.9% win rate on large document analysis tasks and a 78.2% win rate on more advanced web research tasks.
The output side changed to match. You can now export a notebook's results as PNG or SVG charts, PDFs, DOCX or Markdown documents, XLSX spreadsheets, PPTX slide decks, or structured CSV and JSON data, instead of being stuck with a wall of generated text you then have to reformat yourself. There is also a new ability for the tool to search the open web for relevant, credible sources and pull them into your notebook on its own, blurring the line between "tool that reads what you give it" and "tool that goes and finds things for you."
What the free plan actually lets you do
This is the part that matters most if you are not paying for anything. The free tier gives you 100 notebooks total and 50 sources per notebook, with each individual source capped at 500,000 words or 200MB. Audio Overviews and Video Overviews, the feature that turns your documents into a narrated podcast-style summary or a slide-narrated video, are both available on the free plan now too.
The catch is that everything, chat responses, Audio Overviews, and Video Overviews, draws from the same shared usage budget, and that budget resets on a rolling basis rather than giving you a clean daily count to track. Paid tiers raise both numbers: Google AI Pro bumps you to 300 sources per notebook with a larger usage budget, and the Ultra tier goes as high as 600 sources with several times the budget of Pro. For most personal use, researching one topic, digesting a set of PDFs, prepping for one meeting, 50 sources per notebook is rarely the binding constraint. It becomes a real limit if you are trying to run one giant notebook as a permanent knowledge base rather than a project-specific workspace.
Where it beats ChatGPT and Perplexity, and where it does not
The honest comparison depends entirely on what you are trying to do. Gemini Notebook is built around sources you provide, it cites your uploaded documents directly rather than crawling the open web the way Perplexity does. That makes it the stronger choice when you already have the material, a set of PDFs, research papers, or meeting transcripts, and you want answers grounded strictly in that material rather than general web knowledge.
- Perplexity is still the better pick for live, open-web research with visible, clickable citations as you go.
- ChatGPT is the stronger option once research needs to turn into broader work: drafting, spreadsheets, or multi-step tasks that go beyond analyzing a fixed set of documents.
- Gemini Notebook wins specifically when the job is "digest what I already have and tell me what it says," especially now that it can turn that digest into a chart or a slide deck instead of just a paragraph.
A workflow that is becoming common is using more than one of these at once: Perplexity or the open web for discovery, then dropping what you found into Gemini Notebook to actually analyze and cite it against your own material.
The takeaway
A name change alone would not be worth a blog post. What makes Gemini Notebook worth a second look, even if you tried NotebookLM months ago and moved on, is that the free tier now includes features that used to feel like premium territory: code execution, Audio and Video Overviews, and real export formats, all without a credit card. The 50-source cap and the shared usage budget are the honest limits to know going in, not a dealbreaker, just the thing that decides whether this replaces a tool you already pay for or just becomes a second one you keep open in another tab.