Skip to main content
Back

Google's NotebookLM Turns Studying Into a Sherlock Holmes Mystery Game

Google's NotebookLM team published a Sherlock Holmes themed notebook that reframes studying as detective work, showing how persona prompting can transform passive review into an interactive deduction game.

  • NotebookLM published a Sherlock Holmes notebook that gamifies studying as detective work.
  • It uses NotebookLM's public sharing system, so any Google account can open and chat with it.
  • Viewers can chat, deduce clues, and play pre-built Audio Overviews, but cannot add sources.
  • Demonstrates persona prompting on top of existing flashcards, quizzes and Learning Guide features.
  • Follows 140,000+ public notebooks created in the first four weeks of public sharing.
  • Free to use, no install, useful for material that rewards interrogation over recitation.

The team behind NotebookLM just published a new example notebook that turns studying into a Sherlock Holmes mystery. Instead of summarizing or quizzing you on your sources, the notebook role-plays as the famous detective and prompts you to deduce facts, hunt for clues, and reason your way through whatever material you (or the notebook's creator) loaded into it.

It is shipped as a public, shareable notebook rather than a new product feature, which makes it a useful demonstration of how far you can stretch NotebookLM's chat and Studio panel with prompt design alone.

What is actually new

The Sherlock notebook lives in NotebookLM's public-sharing system. Any notebook can be made public and shared with anyone who has a Google account, which Google pitches as a way to make the tool more useful for exploration, understanding and creation. When you open a shared notebook as a viewer, you can read the original sources, ask questions in chat, or consume pre-built artifacts like Audio Overviews and Study Guides, but you cannot generate your own artifacts or add your own sources; the model only answers from the materials the creator provided.

That constraint is what makes the Sherlock framing interesting. The notebook's owner has clearly seeded chat behavior and Studio outputs so the AI stays in character as Holmes, treating the loaded documents as case files instead of textbook chapters.

How to try it

  • Open the Sherlock Holmes notebook link while signed in to a Google account.
  • Use chat to ask Holmes to present a mystery, walk you through clues, or test your deductions on the source material.
  • Play the pre-generated Audio Overview or Study Guide from the Studio panel for a more passive run-through.

There is no separate signup, no paid tier, and nothing to install. NotebookLM itself is free for personal Google accounts, with paid plans only unlocking higher limits.

Why this matters beyond a gimmick

NotebookLM has been pushing hard on active learning. Recent releases added flashcards and quizzes generated from your documents, plus upgraded reports that can produce tailored outputs like blog posts, and a Learning Guide that breaks problems down step by step and adapts explanations, essentially acting as a tutor grounded in your sources. The Sherlock notebook is the next step in that direction: instead of a tutor, a character.

It also rides on the featured-notebooks strategy Google has been building out. After public sharing launched, more than 140,000 public notebooks were created in the first four weeks, and Google has been partnering with publishers and authors to seed curated examples. A Holmes-themed study game is a low-stakes way to show creators what a persona-driven notebook can feel like.

When it is worth your time

This is useful if you are studying material that benefits from interrogation rather than recitation, such as case law, medical differentials, historical sources, or any dense document where the goal is connecting facts rather than memorizing them. It is less useful when you need straightforward recall, where the existing flashcard feature pulls only from your uploaded sources, keeping every card directly tied to your material, is a better fit.

The broader takeaway is small but practical: a well-written system prompt plus a public share link is now enough to publish a working study game, no code, no app, no separate platform.