# notebookLM resource doc

### What is NotebookLM?

NotebookLM is an AI-powered tool designed to help users engage with large datasets of text drawn from diverse sources, such as PDFs, Google Docs, YouTube videos, and websites. By leveraging Retrieval-Augmented Generation (RAG), NotebookLM allows users to query their uploaded materials and receive responses grounded in the provided sources, complete with citations for transparency and accuracy. This RAG framework ensures that generated insights are contextually rich and traceable back to the original content, making it an ideal assistant for research, analysis, and synthesis tasks. Users can also transform complex, multi-source information into digestible formats, such as summaries or conversational outputs.

NotebookLM also offers a the ability to to reflect on and engage with one's own writing and ideas. By uploading past writing, users can ask NotebookLM questions, identify patterns, or even generate podcast-style discussions about their content. This allows for a distance, "uncanny" reflection on one's style, as the AI reinterprets and reframes their content. This not only enhances self-awareness as a writer or thinker but also demonstrates how AI can serve as a partner in ideation and refinement.

### How to get started:

**Getting Started with NotebookLM: A Step-by-Step Guide**

**Accessing NotebookLM**
To begin, navigate to the [NotebookLM website](https://notebooklm.google.com/) and log in using your Google account. Once logged in, you’ll be greeted by an interface of previous notebooks (which will be blank if you are creating your first one). Click the “+ Create new” button to get started with a new notebook.


**Uploading Documents**
The first step in creating a notebook is uploading the materials you want to work with. NotebookLM offers several ways to do this, catering to various file types and sources:

1. Upload files from your computer in PDF, .txt, or Markdown formats.
2. Link directly to a website by entering its URL.
3. Copy and paste text directly into the notebook.
4. Link to a Google Doc.
5. Link to a Google Slideshow.

**Considerations for Uploads**
- You can upload a maximum of 50 files, with each file containing up to 500,000 words.
- You can always upload new documents to a notebook if you stay below this limit. You do not need to add all the document initially.
- NotebookLM currently focuses on text documents, so file types like Excel are unsupported.
- Larger datasets may require additional processing time, so prioritize key documents when necessary.


---

**Navigating the Notebook Guide**

Once you finish uploading your first batch of sources, you will get to the main interface. This is split into three different main windows: *Sources*, *Chat*, and *Studio*.

**Sources:**

In "*Sources*" you can upload new sources, select which you want active, and rename documents. **NOTE**: Renaming documents is very important for later features, as the AI will reference the entire document title, as enter on the sources page. For example, I will rename this PDF of an article from "11.pdf" to "DeepMind Article.pdf"

Do this by hovering to the left of the source title until three dots appear. If you click on these, and click on "More," you can then delete or rename the source.


If you click on a given source, you also get an AI summary of the text, a list of key topics, and the entire text of the source.


---

**Chat:**

When you click on "*Chat*," you are brought to a window where you have a summary of all your sources. Below that summary, you are given the option to add a new note (NotebookLM supports both **written notes** and **AI-generated notes**), generate an audio podcast of the sources (which you can also do from **Studio**), or create a briefing document. This page is also where you can query your documents in natural language. At the bottom of the page, you can find the chat bar, with some example questions pre-populated.

NotebookLM is known in particular for the strength of its Retrieval-Augmented Generation (RAG). All responses include citations, which appear as reference numbers you can click to locate the exact source in your document.


---

**Studio:**

One of NotebookLM’s standout features is its ability to generate **podcasts**. These aren’t simple text-to-speech outputs but dynamic conversations between two AI hosts discussing your documents. You can "Customize" the focus on the podcast, then hit "Generate." You'll have to wait as the tool synthesizes the conversation, which is typically 6–15 minutes long when finished.

Below the podcasts, you can also request that NotebookLM construct "notes." These have various pre-designed formats (Study Guide, FAQ, Timeline, Briefing Doc). You can also save conversations from the chat and source summaries. You can also add notes that you manually write yourself. Saved notes will apear below that.

NotebookLM is also developing a ["video" feature](https://youtu.be/KA_pExdDSUo?si=ri4O9NL7moKUkU9Y), that produces short-form video content with visuals that assist with content comprehension.


---

**Troubleshooting and Support**
As an experimental tool, NotebookLM is constantly evolving. The [support page](https://support.google.com/notebooklm#topic=14287611) is full of information on how to use NotebookLM.

---