Speaker 1: Today, I'm gonna show you how to get DeepSeek R1 to control your computer for free using this powerful locally hosted add-on called WebUI, Browser WebUI. So what you can do is inside the agent settings, you can choose between the agent types, a custom or org. I'm gonna stick with org, then we're gonna go to LM configuration and this is how we're gonna set it up. Now, when you're using this, and I'll show you how to install all of this in a second, you can select Olama. You're not gonna use DeepSeek directly, you're gonna use Olama because that is free and it can be hosted locally. Then you've got the model name, then you've got the base URL. And from here, you can also edit the browser settings. For example, you can say like keep the browser open, window height, window width, et cetera. Then from here, you can run the agent. And basically, this agent then can go off and do anything you want. Now, if we go down to our terminal, we've got two different windows open. As you can see, we have WebUI set up running locally here and then we have DeepSeek R1, the latest version of DeepSeek, the reasoning model that's basically like a O1 competitor but free inside this terminal right here. So you can see I've tested it, it's running the command right there and you can see this is running. So let me show you how to set it up and then let me show you how the computer use agent is powerful and how you can get it to basically run anything. So let's get straight onto this. So first place you're gonna start is by installing browser use WebUI, right? And it's quite simple and easy to set up. So what you're gonna do is copy the commands. If you struggle at any point when you're typing this into terminal like you can see, then what you can actually do is you can actually grab the whole instructions from the GitHub like you see, then I plug it into Claude and I just say like how to install this, right? And for me, for example, it just works perfectly first time and then we've got that running. So that is step number one, all right? And then from here, you can easily just start using web browser, pretty simple and easy to do. Just plug in the LLM configuration and then run the agent. And if you wanna set it up for DeepSeek, then let me show you how to do that, all right? So first thing you're gonna do is make sure you have Ollama installed. You can download it for free. And from here, you can start installing any sort of model you want with AI for free locally. All right, once you've got that installed, then you're gonna install DeepSeek R1. So if you just go search models, type in DeepSeek R1, you can install this model pretty simply. And then from here, we can go directly into a new terminal window like so and just type that command. Now I've already installed it locally, so I'm not gonna do that, but it is pretty amazing that you can install this locally. I mean, if you type in the command Ollama space list, you can see all the models you've installed locally recently. So for example, I've got DeepSeek R1, R1 latest, and QuenCoder 215 latest installed locally so that I can run them anytime. Now what you can do is just type in the command, run DeepSeek latest, and that will start running it for you. So if I say, what model are you? You can see it says, hi, I'm DeepSeek 1, an AI system independently developed by the Chinese company DeepSeek, blah, blah, blah, right? And so we've got WebUI running locally and we've got DeepSeek R1 running locally. Then if you wanna set this up inside the LLM configuration for your web browser, use WebUI. Then what you can do is here, go to Ollama. Then you're gonna select the model name as DeepSeek-R1 latest. The base URL is going to be this. And then once you set it up, you can easily just put in your prompts in your instructions to the agent and then hit run agent. And what that will do is just start running the browser session for you. Pretty simple, you can get all of this set up in minutes. I spent a lot of time researching this so that you could just follow the instructions. If you want the exact step-by-step instructions, I've included it inside the video notes from today, link in the comments description. Just log into the free course, then go to AI agents and then go to DeepSeek R1. So you can follow the exact same commands I use to get this set up so that you can just save a lot of time and just copy and paste. That was super simple and easy to use. Now at this point as well, you might be saying, okay, but is all of this free? So browser use web UI is free. Olama is completely free. Using the LLM configuration that I've shown you for DeepSeek R1 is completely free. If you actually wanna get like the best out of this model, I would actually recommend using Gemini, right? So that is another free API you can use as well. So if you go to aistudio.google.com, create API key, you can actually get the latest new releases from Google AI Studio, but however you want it, you can set this up. And then for example, if you wanna use Gemini, we can say, for example, Gemini Flash Experimental, plug in the API key and we should be good to go. So for example, if we're using Gemini Flash Thinking, it's now analyzing the page. It's Googling OpenAI, it's searching through, it's gonna grab the link in a second. And one of the best things about using the Gemini Flash Experimental API is that you can just, it has vision, right? So you can actually look at the screen, see what's going on and then use that to generate responses. Plus it's pretty fast when you use it. But like I've shown you, you know exactly how to use, for example, Quen, Alarma, whatever you wanna use. And then if we go down to the results section over here, what you can see is it's got the final results. It says the first URL from the search results is openai.com. And we've actually got the screen recording from this session, right? So you can see here how it's analyzing the page with this video. We can watch back the video, see what the agent did. We can actually just do work in the background whilst this is running as well, cause it's on a separate tab. But yeah, super easy and powerful to do. Inside the browser settings as well, you can choose where you wanna record your videos. So when you get the results and the recordings previously, you can choose exactly where they're saved locally. So this is pretty amazing because the number one, you can run R1 locally completely for free using Alarma. Number two, you can have a free AI agent. You don't need to pay, for example, like $200 like I did recently for ChatGPT operator, which is constrained in many ways. Like for example, it's actually blocked from going on YouTube. Number three, you can use any sort of API you want, whether that's DeepSea, Quen, OpenAI, Olarma, whatever you wanna use directly inside the Alarm configuration. And number four, this is all completely free and easy to set up because you can steal all the video notes and instructions directly from me. So let me show you a few use cases for this powerful model. So let's say for example, go to google.com, find the cheapest flights from Bangkok to the UK in July. Hit run agent, that's gonna open up a new tab, completely handless as you can see. Now it's navigating over to Google and it's typed in really quickly actually, cheapest flights from Bangkok to the UK in July. It can actually get rid of the pop-ups. So it can click off the pop-ups as well, which is pretty nice. And bear in mind like ChatGPT operator, which again is $200 a month, that's $2,400 per year. That is constrained so much. Like it can't really use Google Maps and it's actually blocked from using YouTube. And it gets stuck in these endless feedback loops, right? Whereas for example, these sort of APIs, whether you're using Gemini Flash Thinking, which is actually the one I recommend or DeepSeek Reasoning One, et cetera, which is a lot slower because it's obviously thinking out every step like a chess grandmaster in AI. Then for example, when you're using this process, it's much faster, much easier and it can actually do the research you need. So for example, now it's grabbing the flights, it's found the details for us. And the other cool thing about this as well is that you will get a report on the results once this is done. So for example, for research, for analyzing flight prices, it's actually not too bad at all. You'll actually get like an estimation on the task time as well. And I'm gonna just click off the browser right there for now. And then for example, you can see that we can just do stuff in the background, click through the settings, et cetera, whilst we're waiting for that to load. Now, so we've covered the basics now, flights, Googling, whatever, right? Pretty simple stuff, doesn't impress me that much. Now we're gonna try this prompt and you will get access to all the prompts inside the video notes from today. But if we say, go to Llama Coda, direct chat, and start building out an SEO agency website. Be prepared to wait and be patient. We'll just simplify that just for the sake of speed. So simple landing page, one page website. And from here, we're gonna run the agent. Now you can see it's already navigated to Llama Coda. It's analyzing the page. It's prompting the Coda and the AI to start building out that project. And then it's just gonna wait on the page, right? So the good thing about this is it doesn't get trapped in like open loops. It doesn't struggle. It can build out almost whatever you want. We're getting our project, right? So you're basically getting that computer use AI agent to build out other projects for you. And then you've got like a army of AI agents in the background working and use this stuff together. Here's another example. So here we're gonna build out a landing page in a SaaS business. And by the way, if you've never used this tool, it's like a free version of Bolt. Just creates really simple coding projects and you don't need to log in or anything like that. If you ever wanna stop this at any time, if you wanna stop the agent, you can just click stop and I'll stop anything that's running in the background. Here I've actually instructed it to use a low quality option just because that'd be faster for coding as well. So something to bear in mind there. And here we have a to-do list app built out. Pretty nice there. You could also ask the agent to just download the code for you as well and then actually give it back to you so you can use it to build whatever you want later. But essentially you can get this to just do whatever you want, right? Pretty wild. The AI can build apps with AI, code openly, create something quite nice. Let's test if that works. We'll hit add here and you can see the app actually works perfectly. So simple apps, simple websites, landing pages, et cetera. This agent can do directly. Let's test something else now. So I'm gonna say, okay, go to YouTube, do a competitor analysis on Julian Goldie SEO. Hit run agent and let's see if it can openly explore YouTube, right? So it's typing into YouTube now and then it's looking through the results. Funny that Udemy are actually advertising my content. It's having a cheeky gander through the shorts. Must be a TikTok lover. And then if we go back to results, now that's done. One thing I recommend is asking for a report back. Otherwise just being transparent with you. It literally just said, I have completed the competitor analysis but it didn't give me any information. So just something to bear in mind there. If you just make sure it actually gives you a report back on all the details. So we'll test that again. And now if we go back to results, we scroll down. It's giving us a little breakdown there. This is our competitor analysis report. Julian Goldie SEO subscribers. Here's what the channel's about. Here's his latest videos. And here's the content focus. Actually what's interesting is it realized that I was including a link to a free SEO strategy session in each video as well. And you can actually see here like the thought process behind it. Plus the actions taken, right? So you can see all the details along with the video recording too. So pretty awesome stuff there. That's basically how you can set up a deep seek R1 AI agent completely for free. Host it locally, get it to use your browser and control your whole laptop. The other thing to note here is if we go into the terminal you don't just get computer use agent but you can actually use deep seek R1 complete directly on local as well. So for example, if we were like give me the code for an SEO cost calculator it's going to start going off building that directly inside terminal using OLA. The final thing that I was going to say was that another way to get free API is to go to cluster AI and they have an offer where you can get a hundred dollars worth of free credits. So all you have to do is verify your email address and you'll get a hundred dollars worth of free credits. Once you've done that, then you can go to playground go to deep seek R1 and then you can use deep seek R1 via the API. Now, if you need an API key you just scroll down to API keys, issue a new API key and you can get a complete free one. Some people are going to say, is it actually free, et cetera. You can see, I did this just an hour ago, right? So I got a hundred dollars worth of free credits using the API. Then you can plug that directly into browser use web UI. If you want to know the right settings for that basically you would use the open AI dropdown. So you go to LLM configuration, open AI then inside here, you're going to type in deep seek hyphen AI forward slash deep seek hyphen R1. And then that's the end point that you would put inside the base URL right there. So thanks so much for watching. For all the video notes from today including all the instructions on how to set up web UI Olama, the LLM configuration, et cetera. Plus a step-by-step recipe on how to use all of this and some of my best prompts. You can get all the video notes from today inside my free SEO course, link in the comments description. Go to AI agents, then deep seek R1 agents and you can get access to all the video notes from today. And inside this free SEO course there's tons of amazing stuff including keyword research, link building topical maps, EAT, traffic diversification AI SEO systems and AI agents. So it's pretty amazing what you get in there. There's over 225 lessons. And additionally, if you want to get a free one-to-one SEO strategy session, feel free to book them. We'll show you how we take websites from zero to 145,000 business month and generate hundreds of thousands of dollars in sales and autopilot on this free link building acceleration session. You'll get a free SEO domination plan to discover the secrets of SEO link building. Or answer any questions you have. You'll learn the best link building strategy for your website, plus how to quickly outrank your competitors in link building and how to turn SEO traffic based on what's working for us. Thanks for watching. Appreciate it. Bye-bye.