Gemini Notebook Adds Free Textbooks From OpenStax And S Chand

Google has widened what students can feed into Gemini Notebook, its AI research and study tool built on Gemini. The company said on October 29, 2026, that it is adding hundreds of textbooks from providers in the United States and India, including OpenStax, the Open Textbook Library and S Chand Publishing, at no charge. The addition builds on Expert Intelligence, a cross-Google initiative launched the previous month that lets people work with trusted sources inside Google AI products, and on the existing option to add eligible ebooks bought through Google Play Books.

Once a student adds a textbook, Gemini Notebook grounds every answer in that specific text, with inline citations and direct links back to the original passage. The tool can also turn the textbook content into study guides, practice quizzes and flashcards tailored to the material, rather than generic study aids pulled from the open web. Students reach the catalog by selecting “Add sources” inside a notebook, then choosing “Textbooks” to browse what is available.

The numbers

Google did not publish a precise title count, describing the catalog only as “hundreds” of textbooks drawn from three named providers across two countries. That vagueness is itself worth noting for anyone deciding whether to rely on the feature for a specific course: a course assigned an OpenStax biology text is covered, but a course assigned a textbook from a publisher outside those three names is not, and Google gives no public list a student or instructor can check in advance. The free access also sits beside Gemini Notebook’s own usage limits, introduced in an earlier update, so a heavy user who exhausts a quota still faces the service’s existing caps even with a textbook that cost nothing to add.

For a reader weighing study tools, the practical change is less about price and more about where the authoritative text lives. A notebook answer traceable to a specific paragraph of an assigned textbook is a different kind of output than an answer drawn from wherever a general model’s training data happened to touch the subject, and that difference matters most in courses where instructors already distrust unlinked AI citations. Whether OpenStax and S Chand expand that catalog, or whether other publishers join later, will decide how many actual class reading lists this covers rather than how impressive the headline count sounds.