YouTube is full of great educational content, it’s my go-to for watching tutorials and any useful talks. But it’s easy to forget what I watched after closing the tab, and then it’s just onto the next video. I wanted a practical way to retain the information, at least the useful bits like methods, examples, and speakers’ takeaways. So I started using NotebookLM to turn my watched YouTube videos and playlists into a personal knowledge base.

After seeing the success my colleagues had pairing YouTube with NotebookLM, I had to try it myself. I’ve also previously used NotebookLM to help me learn a new graphics tool, and the bulk of my sources were YouTube tutorials, so I already knew how well it works. The goal was to get a less cluttered watch history and more usable material that I could actually reference when needed.

Why build a knowledge base with YouTube videos?

Making YouTube learning actually stick

There are many reasons someone might want to create a knowledge base from their YouTube playlists. The most obvious being that it's a quick way to get the content, which is helpful for anyone who doesn’t have a lot of time or perhaps struggles with focusing for longer periods. But my reason for doing this is that I simply can’t remember most of the stuff I watched. The only way my brain retains information is through old-fashioned sitting down and note-taking.

There’s a ton to learn from YouTube; the platform is packed with so many tutorials, case studies, lectures, and talks. But because there’s so much of it, it’s near- impossible for me to retain or connect what I’ve learned. I end up treating most of it like content for entertainment rather than education, and just flip through videos like pages in a book. Creating playlists was my attempt to curate a YouTube knowledge base, but they’d just sit untouched most of the time. So if I actually wanted to digest all of this content, I was going to have to come up with a smarter solution, which was NotebookLM.

Setting up my YouTube knowledge base in NotebookLM

Extensions make it much easier

I didn’t want this to turn into yet another content hub that overwhelmed me, so I had to set up an efficient system for getting the right YouTube videos into NotebookLM. First, I had to decide which videos were worth my time, which took a while because I had over 60 playlists. I started by getting rid of the oldest playlists I had forgotten about, then stripping the rest down to 50 videos or fewer (NotebookLM only allows up to 50 sources per notebook). Looking back, creating some new YouTube playlists specifically for the videos you intend to migrate to NotebookLM might make the process more streamlined.

Then it was time to get the content into NotebookLM. Adding the links individually wasn’t an option - there were too many. So I had to rely on something that let me do it in batches. Thankfully, there are NotebookLM browser extensions that make this possible. My favorite is YouTube to NotebookLM because it lets you add entire playlists to a notebook in one go, plus it works on my Brave browser. I ended up creating several notebooks, including one for Figma, another for UX learning, psychology & mental health, a miscellaneous one for more entertainment-leaning videos, and a couple more.

Putting my new YouTube knowledge base to use

Turning my playlists into learning material

How you actually use your YouTube-NotebookLM notebooks will be entirely up to you. I recommend checking out these handy prompts to help get you started. Personally, I start by prompting for basic summaries to get an idea of what I’m actually working with. And if I’m really not up for reading a long list of bullet points, I’ll tell NotebookLM to “keep it under 10 bullets” or something similar. Now I get the highlights in minutes as opposed to watching videos for hours.

From there, I can prompt NotebookLM to guide me to the best starting point. Something along these lines:

Based on the selected sources, create a logical order to learn [topic] from beginner to intermediate level.

Or I can ask for its insights on how palpable the content is:

Which of the selected sources gives the most comprehensive explanation of [topic]?

Or have it pick out the top videos to start watching:

Act as my personal study coach. Pick the first three videos I should watch, and explain why they’re a good starting point.

Clicking on the cited references takes me straight to the video. What I love about NotebookLM is how it handles YouTube videos; you can watch them right in the source panel, while taking a Note on the other side, and prompting the AI if you have questions along the way. This is the ultimate learn-as-you-watch setup for me - it’s a video player, AI assistant, and note-taker all in one, and there’s nothing to distract me from learning. I’ll also take notes of the videos I’ve “completed” so I know to deselect them when prompting NotebookLM for the next steps.

For the notebooks that don’t follow a clear learning progression and don’t have beginner-to-expert levels, like my health resources or videos about prehistoric animals, I take a broader approach. I’ll ask for summaries of the topics, take my pick, and dive deeper from there. This is also a great way to extract common themes or conflicting viewpoints across the videos. For example:

What are the most common habits mentioned in these videos?

And this is what I use if I want to cover all my bases and get all the different angles on a specific topic:

Extract conflicting advice, opinions, or conclusions across these videos about [topic]. Summarize what each creator argues.

From watching to understanding

Creating a YouTube knowledge base in NotebookLM completely changed how I actually interact with my YouTube videos. Instead of watching in the hopes of gaining something, I’m now actually learning from these videos and often get additional insights and connections. My favorite part is how everything is organized into one workspace: video + AI chat + note-taking. Now when I find an educational video that may be useful, I’ll add it to the relevant notebook. And if a notebook starts getting full, I can either create a new one or just prompt NotebookLM to tell me which videos are worth keeping or removing based on what I’ve already learned.