Google adds Video Overviews to NotebookLM

NotebookLM's viral AI podcast hosts are finally getting some visuals to go with their chatter.

On Tuesday, Google officially began rolling out Video Overviews to NotebookLM, its research and note-taking assistant. First teased at Google I/O in May, the new feature lets you turn raw notes, dense PDFs, and documents into bite-sized, narrated video presentations.

Instead of just listening to AI hosts riff on your uploaded material, Video Overviews acts as a visual alternative. The AI generates slide-style visuals while pulling exact charts, quotes, diagrams, and figures directly from your source files. It is designed to tackle the kind of dry data and abstract processes that are tough to grasp through audio alone.

Custom pitches and a much-needed Studio overhaul

Just like the Audio Overviews, the video output is fully customizable. You can prompt the AI to treat you like a complete beginner who just wants the diagrams explained, or tell it you are already a specialist and want the focus dialed into a specific section.

Along with the video feature, Google overhauled the Studio panel to fix one of NotebookLM's biggest annoyances: the single-output limit. You can now create and save multiple versions of the same asset inside a single notebook.

That means you can spin up different audio guides for different languages, tailor video briefs for separate teams, or generate separate mind maps for each chapter of a study guide. The panel now features four dedicated tiles at the top for Audio Overviews, Video Overviews, Mind Maps, and Reports, and allows multitasking so you can listen to a podcast while clicking through a study sheet.

Why it matters

NotebookLM struck gold with Audio Overviews because people love passive consumption, but listening to two AI voices explain complex charts without visuals always felt half-baked. Adding narrated slides bridges the gap between a dense paper and an intuitive briefing, making it actually usable for complex visual topics.

We think the real workhorse upgrade here is unlocking multiple outputs per notebook. Previously, generating a new study guide or podcast meant wiping out your previous one, which essentially made it a parlor trick rather than a serious research workflow. Letting users store multiple tailored decks and cross-reference them while listening makes NotebookLM feel like a full-fledged collaborative workspace rather than just another LLM sandbox.