Gemini Notebook Adds Real-Time Voice, Video Lessons and AI Testing

Google is expanding Gemini Notebook as a mobile-first AI research and study application, adding real-time conversational voice capabilities in more than 100 languages, automated short-form video generation, and diagnostic assessment tools. Google Director of Product Management Trond Wuellner detailed the update, which moves the platform beyond document summarization toward an interactive audio-visual study interface.

The application was created within Google Labs to address large language model “hallucinations,” in which generative AI invents facts. Its architecture restricts responses to files, transcripts, PDFs, and links uploaded by the user, rather than drawing generic information from the open internet. The primary voice and recording tools are launching on mobile devices before expanding to web and desktop platforms.

NotebookLM began as a prototype called “Project Tailwind” at the Google I/O developer conference in May 2023. It later launched publicly as NotebookLM and was eventually integrated into Google’s Gemini ecosystem. The mobile applications for iOS and Android now include an audio recording tool for live lectures, personal spoken notes, and ambient voice thoughts.

Those recordings are automatically transcribed and stored alongside photographs of physical notes, handwritten documents, and digital text files. During study queries, the system can reference and cite exact timestamps from the audio. Users can also hold spoken conversations with their uploaded study notebooks in real time, interrupting the spoken output to ask follow-up questions, clarify specific terms, or change the direction of a study session.

The voice interaction engine remains grounded in the user’s uploaded repository, keeping responses tied to the provided coursework. The update builds on the previous “Audio Overviews” feature, which generated synthetic two-host podcast discussions summarizing uploaded documents.

Gemini Notebook now also generates automated 60-second educational videos. These short visual clips combine spoken narration with dynamic animations illustrating complex mathematical formulas, scientific diagrams, and structural concepts. Users can save the 60-second video summaries and share them with classmates or research partners.

Google has revised the study reports section to bring written summaries, infographics, flashcards, and interactive practice tests into unified review dashboards. The assessment system goes beyond basic flashcards with multiple-choice questions, short-answer prompts, and fill-in-the-blank tests.

After users complete a quiz or flashcard module, they can ask the AI to evaluate their performance. The system examines correct and incorrect answers to identify subject areas that require further review. Users can manually edit, add, or remove quiz questions so they match specific course curricula.

According to Google, real-time voice chat and live audio capture are immediately available on iOS and Android. The full suite of interactive reports, video creation tools, and custom testing features will roll out gradually across web and desktop versions.