Google adds TikTok-style video generation from user notes to NotebookLM

In July 2026, Google added a new summary format to its AI service NotebookLM: short vertical TikTok-style videos that the service generates automatically from documents and notes the user has uploaded.

How the new format works

NotebookLM turns source material such as articles, lecture notes, PDF files and slide decks into a finished vertical video. The AI writes the script, picks the illustrations, and adds subtitles and a synthesized voiceover on its own. The format clearly targets the habits of short-video audiences on TikTok, YouTube Shorts and Instagram Reels, where users absorb a quick rundown of a topic faster than they read a long text or listen to a half-hour podcast.

• The new format: vertical TikTok-style videos that NotebookLM assembles from the user's notes and documents
• The service already supported podcast-style audio summaries and text overviews
• Google pitches the videos as a way to replace "boring lectures and presentations" with short, fast-paced clips

How NotebookLM has evolved since 2023

NotebookLM launched as an experimental Google Labs project built on Gemini models. The service answers questions only from documents the user has uploaded, such as PDFs, articles, video links and notes, and cites where in the material each quote comes from, which reduces the risk of hallucinations compared with a typical chatbot. In 2024, Google added Audio Overviews, which turned uploaded notes into a podcast-style conversation between two virtual hosts. The format unexpectedly went viral: users widely shared clips in which the AI voices sounded like a real, live conversation. TikTok-style video is the logical next step for the same idea: the same source content, in the format social media audiences consume fastest.

Why Google is changing the summary format

Google originally designed NotebookLM as a tool for working with notes and research material: building topic maps, answering questions from sources, and producing audio overviews. Video adds another way to package the same content for different consumption habits. Some people prefer to read, some to listen on the go, and some to watch a one-minute clip instead of an hour-long lecture.

Who the format is for

The most obvious audience is students, who may find it easier to remember material from a short, fast-paced clip than to reread their notes before an exam. The format also suits teachers who need quick topic overviews for class, and teams that need to explain the gist of a long report or study to colleagues in a minute without building a separate presentation. Because the video is generated automatically from sources already uploaded, users don't have to write a script, edit video or pick illustrations themselves; the model handles all of that.

What it means

Google keeps building generative AI into everyday education and work. Instead of a student or employee assembling a presentation from their notes, the model does it for them, and in a format people are more likely to watch to the end. For NotebookLM, it's another step from a niche note-taking tool toward a full-fledged AI content studio.