I opened up my favorite research tool last week and half of it looked different. The name changed, the buttons moved, and a few features I never noticed before were suddenly front and center. Then I found this walkthrough from the team over at Futurepedia, and the creator broke down every single update to what used to be called NotebookLM, now Gemini Notebook. I was genuinely surprised by how much more powerful it got.
If you have never touched it, here is the one-line pitch the author gives: it is an AI that digs into a specific topic using only the sources you approve. That means way fewer hallucinations, because it is grounded in your material. You just log in with your Google account at notebook.google.com and go.
What’s new
The rename is the headline, but the real story is under the hood. According to the walkthrough, three panels now run the show:
- Left panel (your knowledge base): the sources you upload or let Gemini find and approve.
- Middle panel (chat): feels like a normal chat, but it only pulls from your sources and drops inline citations you can click to read the exact context.
- Right panel (studio): where you turn all that research into podcasts, slide decks, infographics, videos, quizzes, and more.
The creator points out that sources can be almost anything now: PDFs, websites, YouTube links, audio, images, even living Google Drive docs that update automatically when you edit them.
The twist
Here is the part that got me. Gemini Notebook is famous for staying locked to your sources. But the expert shows a new behavior most people miss. If you ask it something like “are there contrarian or lesser-known viewpoints not represented in these sources,” it understands it is allowed to break out, think, and go hunt down fresh reputable sources. Then the moment you ask a normal question, it snaps right back to its careful, source-only mode.
So it is both a walled garden and a researcher, depending on how you phrase the ask. That flexibility is the thing I think changes how you use the whole tool.
The step-by-step workflow
Here is the mini-workflow the original poster runs for a tricky topic (their example was the environmental impact of AI):
- 📥 Load your sources. Upload PDFs, paste website links (hit space between each), drop in YouTube videos. If a source turns red, download it as a PDF or paste the raw text instead.
- 🔎 Let Gemini fill the gaps. Type your topic and let it search the web for more sources. The author strongly recommends verifying each one comes from a reputable site before approving.
- 🧠 Run three analysis prompts. This is the creator’s favorite trick for nuanced topics:
- Ask for contrarian or lesser-known viewpoints missing from your sources.
- Ask what important subtopics or questions are barely covered.
- Ask where the sources disagree or contradict each other.
- 💾 Save the good answers. Click “save to note,” then “convert to source” so Gemini keeps that nuance in mind as it responds later.
- 🎨 Transform in the studio. Pick a format and generate.
That third step is where the real gold is. The expert found a wild claim buried in the research: using AI for writing and illustration produced 130 to 2,900 times lower carbon emissions than human writers and illustrators. Even the creator admitted that one was strange to sit with. Surfacing conflicts like that is exactly why those prompts matter.
The studio panel is the star
This is where Gemini Notebook pulls ahead of everything else, and the walkthrough demos each format:
- Audio overview: a podcast-style deep dive you can even jump into and talk to live.
- Slide decks: presenter style or detailed, with a revise button for edits.
- Video overviews: cinematic, explainer, or the new short-form “educational doom scrolling” clips.
- Mind maps, reports, flashcards, quizzes, data tables: great for retention or polished deliverables.
- Infographics: genuinely social-media ready.
The person who made the video is honest about the rough edges too. The cinematic videos take forever and often are not good enough end to end. The “detailed” infographic setting crams too much in and spits out spelling and logic errors.
Pro tips worth stealing
A few gems the creator shared that I would not have figured out alone:
- Fix messy infographics fast. Drop the exported image into ChatGPT and ask it to fix all errors. The author says GPT image tools catch misspellings, bad grammar, and even wrong equations, often coming back perfect.
- Use auto-label to focus. Let it sort sources into topic tabs, then deselect whole groups so the chat answers from just the sources you care about.
- Clean up your notebook. Rename the auto-generated titles and summaries to something short and literal so you can tell notebooks apart at a glance.
- Bounce to Gemini for creativity. Every notebook now shows up in the Gemini app, where you can run multiple separate chats against the same knowledge base. Accuracy in Notebook, creativity in Gemini, and they sync both ways.
I love that last one. The tool’s biggest strength, its accuracy, can feel limiting when you want to brainstorm, and this fixes it neatly.
The creator also packed all of this, including the three analysis prompts and five step-by-step use cases, into a free written guide linked with the video. If you want to actually master Gemini Notebook instead of just poking at it, watch the full walkthrough from Futurepedia below and grab that companion resource. It is the fastest way to see every one of these features in action. 🚀