[0:00]What if your entire research stack, Perplexity, every expensive platform, every subscription you signed up for during a 3:00 a.m. productivity spiral could be replaced with two free tools and a Python library some random guy built in his spare time and posted on GitHub?
[0:17]Because it actually can, and it's low key embarrassing how long this has been sitting in front of everyone.
[0:22]One of the tools you've definitely heard of, you tried it once, thought it was cute, maybe made one podcast with it, and then completely ghosted it by Tuesday. Notebook, LM.
[0:32]Yeah, that one. Turns out nobody's been using it right, including you.
[0:36]The second you wire it into Claude code as a rag backend, it stops being that app collecting digital dust and becomes the most powerful free research agent you've ever touched.
[0:45]No context tokens getting nuked on heavy analysis, no $200 platforms, no subscriptions, just two free tools doing something that genuinely should not be legal at this price point, which is zero.
[0:58]In this video, I'm showing you exactly how to build it.
[1:01]So, I got anti-gravity here open with Claude code and I just want to show you guys how powerful this is with a quick demo.
[1:09]So, first, I just said add the YouTube search skill.
[1:13]That's just going to bring the context in and then I said, use the YT search skill to find the latest trending videos on Meta Ads strategies and tactics.
[1:23]Once we have those videos, send them over to Notebook LM using the Notebook LM skill.
[1:28]Analyze the top Meta ad strategies, frameworks, and creative trends discussed in those videos.
[1:34]Then have Notebook LM generate an infographic in a hand-drawn/blueprint style that visually summarizes the key Meta ads insights, strategies, and winning approaches.
[1:46]So now I'm just going to send that off and it's going to start doing all the automations in the background and I'll show you guys how it comes back.
[1:55]All right, and it just finished running. As you can see, it gave us a super comprehensive analysis that's up to date with all the new trends and frameworks for Meta ads.
[2:09]And what's nice about this is that it's based on the top 20 viral videos about this specific topic.
[2:19]So, it has really quality information. Not only did it give us this comprehensive analysis, but it also created the infographic just like we asked it to.
[2:28]And if we actually take a look here, on the left, we have the Meta Ads Blueprint infographic. It actually downloaded it and brought it straight into our project.
[2:37]Here you can see that it actually used the style that we asked it to use.
[2:45]And this is using Nano Banana Pro, so all this information is correctly spelled and up to date based on all the information that we extracted from those videos.
[2:56]All right, so now I'm going to show you guys how to actually set this up so that you can connect Notebook LM with your Claude code and you can do some sort of similar workflow for yourself.
[3:07]The reason I'd also do this is because Claude code isn't the best at doing research.
[3:14]So, if you want up-to-date information on any kind of topic, you can just use this workflow to get as much research as you want.
[3:24]Also, since it's using Notebook LM, you're not using up your Claude code tokens to do the research.
[3:30]It's virtually free research because you're using Notebook LM.
[3:37]So, you don't have to use any tokens for this. The only tokens is to send the information to Notebook LM and then get the information back.
[3:42]So, why this is so powerful is because if you're trying to build a SAS or you're trying to build an agentic workflow, well, now you can import all this different information from different resources, like the top YouTube videos for that specific niche.
[3:57]And you can just create this ultimate workflow that has the most recent and up-to-date information on how to do all the things together.
[4:05]So, now onto the fun stuff, how to set this up. So, the first thing it does is basically we get all our resources from YouTube.
[4:13]So, how we do this is we actually have a skill here called YouTube search skill setup.
[4:19]And if we actually take a look in the skill, it's just a skill telling Claude code how to use the YT-DLP tool to be able to search YouTube.
[4:31]Now, as you can see, it's quite a comprehensive guide on how to do it and how to get the videos back so that you can use them for Notebook LM.
[4:41]Now, there's two ways to actually get this skill. The first way is you can just go into Claude code and ask it to create you a skill for how to use YT-DLP.
[4:55]And Claude code should be able to cook something up for you.
[4:58]But if you do want this specific skill that I created, then you can just go to the first link in the description to my school.
[5:05]I will have the skill over there. You can just download it and drag and drop it straight into your project.
[5:10]Now, the next thing we'll need is an API connection to Notebook LM.
[5:17]Now, how do we do this? Because Notebook LM doesn't have an API key right now.
[5:22]Well, thanks to Teng Lin right here, he actually created a makeshift API connection script that you can use yourself to connect to Notebook LM, and his repo is public, so you can use it for yourself.
[5:35]So this repo is called Notebook LM-py, and as you can see right here, it's super simple to set up.
[5:43]All we're going to do is we're going to scroll down here and we're going to come to the installation part.
[5:48]Then, you're just going to take this piece of code right here, pip install notebooklm-py.
[5:53]And we're just going to come back to our IDE or your terminal.
[5:58]And you're just going to want to paste that code into your terminal, and once you run that, it's going to install this whole Notebook LM ecosystem that will allow you to connect.
[6:08]Now, there's another thing you have to do once you run that, and that's going to be to connect your login to Notebook LM so it knows which instance of Notebook LM to use.
[6:20]Now, if you scroll down right here, it's going to say the first thing to do right here is authenticate, which is you're going to run this notebook LM login command, and that's going to open a new browser where you're actually able to log in to your Google account and that's going to connect everything super seamlessly.
[6:36]So, just copy this right here. Go back to your terminal, run this, and then you should be able to log in and that's just going to create the connection.
[6:44]Now, there is one more thing we have to do to make this work perfectly fine.
[6:50]If you scroll down lower, you're going to see here that he actually made a agent skill for Claude Code, and to install this, we're just going to use the Notebook skill install command.
[7:00]So, once again, just copy that, go back to your terminal, and just paste that right in, run that as well.
[7:06]And that's going to download the script that tells antigravity or Claude code how to use the script that he created so that it all works perfectly fine.
[7:16]Now, once you did all that, your setup is complete, you can just open a new instance of Claude code in that project, and you can start asking it to create you a notebook on any research topic, and then you can ask for the analysis, you can ask for an infographic, you can ask for basically anything.
[7:35]So, it can actually even do things beyond the web UI of Notebook LM that only this kind of setup can do.
[7:44]So, for example, we can do batch downloads, quiz/flashcard export, mind map data extraction, data table CSV export.
[7:54]We can do slide decks as a PowerPoint file. This is very, very useful.
[8:00]We can do slide revisions with natural language, save chat to notes, source full text access, and programmatic sharing.
[8:08]So I hope you can see how useful this whole ecosystem is.
[8:11]This actually connects to your Notebook LM and just kind of uses like an API to do all this stuff.
[8:18]So, if you go to Notebook LM, you can see right here, it created me this notebook, and all the videos that it found are right here in the sources, as well as the analysis right here.
[8:29]And over here, you can see the infographic that it created.
[8:33]Just like that. All right, so I hope you did enjoy this video.
[8:37]If you did, please leave me a like and a subscribe. It's greatly appreciated since I am a smaller channel and I'm trying to grow.
[8:44]Also, if you do want all my N template and skills, as well as all my resources on how to use AI the best to actually be able to make money with it.
[8:55]Well, then you can join my school, first link in the description down below.
[8:59]If you're a business owner that needs help with AI or marketing, you can check out my agency as well and book a call.
[9:07]But either way, guys, I hope you had a great day and see y'all in the next one.
[9:12]Peace.
