Speaker 1: This past week has been absolutely insane with the release of the new DeepSeek R1 model, the most powerful open source model ever created. It's already outshining proprietary models like Cloud 3.5 Sonnet and GPT-4 Omni across nearly every benchmark test. Developers everywhere are actually pulling all-nighters integrating this groundbreaking model into their dev tools such as Bolt, vZero, Cursor, and so many others. But here's the catch. A lot of these tools sit behind paywalls, which might not work for everyone. But don't worry, I have a solution to this. Today, I'm going to showcase how you can harness the power of the DeepSeek R1 model with a familiar tool that we've showcased multiple times on this channel called Bolt.diy. Now unlike many paid tools, Bolt.diy is completely open source and driven by a community led by an amazing YouTuber called Cole Medin, who continuously delivers updates with the community to make the platform even better. Now for the people who haven't heard of Bolt.diy, let me explain. It's a platform that allows you to prompt, run, edit, and deploy full stack applications on the web using any LLM of your choice. You can create functional websites and interactive pages, animations, and so much more. All from just a few prompts, without writing any single line of code. Here is a quick example of what you can do with Bolt.diy. This is where someone had created this fully functional website with Bolt.diy, and it's essentially something that was created with a few prompts. You can see that it has animations, functional chatbots, you have the ability to have it so that it could code out full on pages, and this is the capability that you get with Bolt.diy. I've made multiple videos showcasing Bolt.diy, so I definitely recommend that you take a look at it with the links in the description below. But now, let's take this to the next level and have it combined with the DeepSeek R1 model, where we can get its reasoning capabilities and thinking process to accurately build out the best full stack application that's available. So with that thought, let's get started and showcase how you can fully set this up locally. And the great thing is guys, this is completely for free. You can either install this locally with Olama, or you can actually get a free API key from Hyperbolic, where you have up to $10 worth of credits that you can get completely for free for the DeepSeek R1. And in this case, you can go ahead and copy their API key, and then you can paste it in later, which I'll showcase afterwards into Bolt.diy. Now to install this, you have a couple of different methods. Now if you're going to be following through with my installation process, I'm going to be directly installing it. So I'm going to be following through with these following commands, but you can also use Docker to install this. And they also have a source file that you can install to quickly get started. But if you're going to be following through with option one, you're going to need to fulfill the prerequisites. You need to make sure that you have git installed to help you clone the repository locally. You'll need to make sure that you have node.js, as well as the pnpm command to install the certain packages. So once you have these prerequisites fulfilled, go over to this repository, click on this green button at the top, and copy this link to clipboard. Now you can scroll all the way back down to the installation demo, and you want to then simply open up command prompt. Once you have opened up command prompt, type in git clone, and then paste in the link and click enter. Now I'm getting this error because I already have bolt.diy cloned onto my computer, but for the people who are going to be doing this for the first time, it's going to start cloning this into the desired path that you have specified. But once it has finished cloning, you can simply type in cd bolt.diy to get into the bolt.diy directory. Then you want to go ahead and install the packages with the npm command. So go ahead and paste this in. This will start installing everything that is necessary. Once that is done, install the project dependencies. So you will need to use the pnpm command to install all that. Once that is done, you can then start up the application with the pnpm rundev command. So go ahead and paste this into your command prompt, and this will start this up within your local host so that you can then access it on the web. So there we go. We can then go ahead and copy this. This will take a couple of minutes to start up depending on your computer, and essentially you're going to be then able to access it through the local host after it loads up. And there we go. We have bolt.diy locally installed with all the new updates, and now we can basically get started and implement the new DeepSeek R1 model. Before we get started, I got a huge new update. This is where I've launched a new newsletter. This is something that's going to be sent out on a weekly basis, and it's essentially going to be updating you on the latest AI advancements, comparison of different large language models, AI news, as well as ranking different AI agents. So definitely go ahead and subscribe to this because you don't want to miss out on free AI news. So guys, there's actually three ways to use the new DeepSeek R1 model. You can either provide your DeepSeek API key that's paid and fully available on the cloud through DeepSeek's API and their console, and you can do this within this section over here. You can go ahead and configure and paste in your DeepSeek API key, or you can go ahead and you can use OpenAI compatible to then provide the free hyperbolic API key. Now, the best method, in my opinion, is going to be installing this with Olama so that you can then work with the different distilled models. So once you have the models installed, you will then be able to go over to this drop down menu and then click the model that works best for you, and you're going to need to actually set the API or the .env within this file to then access the model of your choice. Essentially, to install Olama, it's super easy. Go ahead and install it for your operating system. Once you have it installed, you want to then go over to the model card of DeepSeek R1, and then you want to install the different distilled models. Obviously, the 671 billion is going to be the best, but no one's going to be able to run this model, which is why you would want to use this on the cloud API that DeepSeek provides or hyperbolic provides. But you're still going to get great results with the 14 as well as the 32 B parameter model, but it's just going to require you to have a high end PC. Some people have actually messaged me saying that the 7 billion has been doing decent work for them in terms of debugging as well as assisting in basic error messages and error handling. So it truly depends on what computer you have, but select the one that is going to be best for you, and then you can easily get started as you can then easily load it up over here and start generating anything you would like. So let's get started. This is where I had requested to create a tic-tac-toe game in HTML, CSS, and JavaScript only, and it was able to easily build out this game within a couple of seconds, and it was able to generate the code. And now we're going to work towards making more intricate and different types of builds with this new combination of bolt.diy as well as with the new DeepSeek R1 model. So what we're going to do now is have it generate a financial app that works with tracking stocks like tech stocks and where you can manage your portfolio like investments, recent, etc. And right away, it is working on creating the basic front end components. You can see the code is being written, and it is quite quick in doing so. But remember, guys, this is a reasoning model, so it might take longer in certain sections for it to process and think through before generating because it accurately works with different agents within the model to help you process the generation. So let's see what it ends up building out. And there we go. We have the first iteration of this financial portfolio manager. So you have the ability to track your stocks, REITs, as well as ETFs. You can paste in the symbol, and I assume this is where you're supposed to put the quantity amount of stocks you have, as well as the price of the stock. So I've been playing around with the R1 model, and I added a bit more functionality to it. And in this case, I kind of removed the graph because it wasn't working with all of the different stocks. So essentially, this is where you can go ahead, search up a stock. You can then add it to the list, and you can see the daily changes that were made. Right now, obviously, it's not going to show the updated price because it is after 430 right now. But you can see that you're going to be able to add different stocks. You can then even track their individual growth with this graph over here when you click on it. So you can see that it has been able to develop this investment dashboard that lets you track your investments on the go. You can even create this as a web app with the DeepSeek R1 model. So you can see that there's small things like this that could be developed with this model, and it's essentially going to help you a lot in different ways. If you like this video and would love to support the channel, you can consider donating to my channel through the Super Thanks option below. Or you can consider joining our private Discord, where you can access multiple subscriptions to different AI tools for free on a monthly basis, plus daily AI news and exclusive content, plus a lot more. Now, guys, this is truly insane. This is definitely a great combination, and I definitely urge you to try this out because it's going to elevate your performance with code, as it's going to increase your efficiency, productivity, as well as make it so that the code it generates is going to be more effective due to its reasoning capabilities. So definitely try this out with all the links that I have in today's video in the description below. Make sure you follow me on the Patreon. Follow me on the newsletter. You should definitely stay up to date with my daily posts over here with tweets. And lastly, make sure you guys subscribe, turn on notification bell, like this video, and please take a look at our previous videos, because there's a lot of content that you will definitely benefit from. But with that thought, guys, have an amazing day, spread positivity, and I'll see you guys fairly shortly. Peace out, fellas.