NotebookLM turns docs into anime and papercraft videos

Turning dense documents into engaging content used to be NotebookLM's party trick with audio podcasts. Now, Google is giving the visual side a serious makeover.

Google just rolled out a major upgrade to Video Overviews in NotebookLM, powered directly by Nano Banana, the image generation model from Gemini 2.5 Flash that debuted back in August. According to Google, users have already generated over 5 billion images with Nano Banana inside the Gemini app.

Now, Google is plugging that horsepower straight into real-world tools, hitting NotebookLM and Google Search Lens today, with Google Photos on deck.

Six new aesthetics and a bite-sized format

Instead of generic visuals, NotebookLM now creates contextual illustrations tailored to the actual sources you upload.

Users can choose between six distinct artistic styles: Watercolor, Papercraft, Anime, Whiteboard, Retro Print, and Heritage. Whether you want a clean whiteboard summary for a sprint planning doc or an anime vibe for a history study guide, you can set the mood.

Google also split Video Overviews into two distinct formats. Explainer delivers a structured, comprehensive walkthrough for deep dives. If you are in a rush, Brief creates bite-sized videos focused strictly on top-level takeaways.

Customization goes deeper, too. You can tap the pencil icon to give steering prompts, like telling it to focus only on the cost analysis inside a 50-page pitch deck, or to turn a recipe collection into quick cooking step highlights.

The rollout kicks off this week for Pro tier subscribers across all supported languages, with free users gaining access over the coming weeks.

Why it matters

We think this is a smart playbook from Google. While rivals keep image generation locked inside chat prompts, Google is quietly embedding its proven model—one with 5 billion generations under its belt—into functional workflows like Search and NotebookLM.

For students, analysts, and knowledge workers, this lowers the friction of parsing dense materials. Audio podcasts were great for passive listening, but visual diagrams and whiteboard-style recaps make technical arguments stick much faster. It is definitely worth trying as soon as it hits your workspace.

The broader ripple effect is that pure text summarizers are going to look increasingly outdated. Document AI is evolving into a full multi-modal production studio. The main question mark going forward will be factual fidelity: artistic styles like anime or watercolor look fun, but Google will need to make sure stylistic flair never distorts real data, charts, or critical business nuance.