Kimi K3 AI Full COURSE 1 HOUR (Build & Automate Anything)

In this one hour course, I'm going to show you exactly how to use Kimi and Kimi K3 specifically, but this works for any other model. So if you're watching this after Kimi K3 has dropped, don't worry, this will work, it's evergreen. And you might be wondering, okay, why use Kimi? Well, basically, this is one of the most powerful, if not the most powerful model in the world right now.

It's open source, it comes from China, it's cheap to use, and it's very, very powerful. I actually tested out on 50 different builds, which I'll show you in a second, and I'm going to go deep into how to use Kimi across all the different methods, because there's so many ways you can use Kimi that most people don't even realize. So it's a fantastic tool for coding, it's self-improving, it's only getting better. There's been giant leaps forward in terms of the gap between Kimi K2.7 and Kimi K3, and it's only getting better.

So this is one of those powerful tools you can use, but most people don't know how to use it, and that's what we're going to get deep into inside this video. This is by far the best open source model. So I've tested GLM Vibrant 2, tested, for example, Tensor and HY3. This is the best by far.

Now, if you want to get started, you can go to Kimi.com, and we'll start using that in a second. I also want to show you that we've actually tested this out with 50 different builds using K3. So if you're like, okay, you know, what does this guy know about it? We've actually tested it out.

I mean, it only just dropped yesterday, but every single model of Kimi I've already tested personally, and we've built out so many things that you can actually go and check out on GoldieBench. And it is right up there with, for example, frontier models like GPT 5.6 and Fable 5. As of recording this, you can create some beautiful stuff. You can build some amazing things.

And I'm going to show you how we build stuff like this and how you can get the most out of it, and also how you can plug Kimi into coding agents so that you can automate your work 24-7. Also be looking at Kimi Claw, Kimi Work, all sorts of places that 99% of people have never tested out when it comes to Kimi, but they're very, very powerful. So with that, let's get straight into this. We're going to start working with Kimi.com.

Now, Kimi.com, it's kind of like ChatGPT.com, but Kimi.com, you can do all sorts of agentic things you actually can't do inside ChatGPT as well. So you'll see, for example, there's these presets like swarm, slides, deep research, websites, documents, and sheets. So this is what it's going to look like once you land on Kimi.com. Now, you can just prompt it, basically, like so.

So you can type something in like, for example, build me a beautiful website for an SEO agency, as an example. And I just want to show you some examples of how you can use this and how you get started with this. And then what I usually like to do is just run Kimi in parallel in different tabs, right? So you can open it inside your browser and then run it in different tabs.

And you can see what it will start doing is it's like, right, I'm going to build this website. We're going to start using that tool. It's reading our skill MD for creating websites. And then I'm actually going to install Kimi Work in the background.

I think we actually just need to log in with Google if you're setting up Kimi Work, but we'll come on to that in a second. So what you're going to see here is that when it starts building something like a website, it's got the chat on the left-hand side, and then you've got this kind of preview, this computer on the right-hand side. So what it's actually doing is it's got a computer in the cloud. There's a virtual computer, and then you can preview what it's doing over here.

Now, if you want to minimize that, you can just click the minimize section over here, and then you can see an example of what it's doing. Now also, as it's going along, it's going to give you details of what it's working on. So for example, you don't need to be a coder to be able to use Kimi. You can just type what you want, explain it very simply.

And then from there, it will start building that out, and it will handle all the technical stuff for you. So you're at the point now where you don't need to be technical to build anything, and you can build games, apps, websites, tools, whatever you prefer using this. Now, in the meantime, what you'll see here is that we've got a list of models we can use. The model that you really want to be using if you want to get the most out of this stuff is K3 Max.

Now, you'll also see this option for K3 Swarm, which is like a team of agents, like a big, big batch of agents working together, and we'll come on to that in a second. And then we have Kimi K2.6. Now, K3 is the latest model. K2.6 is an older generation.

It replies a lot faster. So if you don't want it to be like agentic like this, so you can see it's a bit slow, it's reasoning through, it's planning, it's working in the background. If you need it to be much faster, and you just want like instant responses, you don't want to wait around, then you can switch over to Kimi K2.6 Fast. And also, you'll see here that we can switch the context window.

Now, on the previous version of Kimi K2.6, you had 256k tokens, and you had extra long, which is a million context window. You might say, what's the difference? Well, basically, there's a four times difference in the context window, which means that you can handle and process four times more. Now, if you're having a really long conversation, or you're giving a lot of data to Kimi, then you'll need extra long.

But if you just want like something basic, like creating a website, and there's not much information that you're feeding to Kimi, and you're not planning to go back and forth with it for hours and hours, then you can just go with the basic standard context length. So we can leave that running in the background. And then what we can do over here is we can switch over to Swarm. And you'll see here that it actually gives us this option for Swarms over here.

Now, it's actually got some examples. If you're wondering, okay, what can you do with Swarms? Well, it's actually got some featured cases over here. So we can actually preview what's been built previously with these models.

So for example, I actually created this black hole simulation, which is absolutely beautiful. It created this amazing, interesting sort of like website. So it's very good at website designs. And the cool thing about this is, well, it doesn't look like AI slop.

It's not just like your typical sort of AI design. And then of course, we have this orbit map, right? And we can have a look and we can see, okay, do you want to show the GPS? Do you want to show the space stations on this orbit?

It's just like a full app. Now, if you see any examples, and this could be for websites, for Kimmy Swarm, whatever, you can click on Remix, and you can actually use the same prompt to run through the same workflow yourself if you want to test it. If you want to remix it, well, when you're remixing, basically what you can do is you can edit the prompt before you create it. So if you click on Remix over here, it gives you the prompt for that original showcase.

And then you can customize it. So you can say, okay, use my brand colors or change the website design to this, or use the same website design, but make sure you add this in, blah, blah, blah. Really cool. Now, if we wanted to test out Kimmy Agent Swarms, we can do an example like this.

So we can say, build out a beautiful website for a blog about Hermes Agent and create five blog posts on that website, do all the keyword research, and then build it out and show in a preview. Use my brand colors and make it look beautiful. And the difference between Swarms and normal Kimmy is that with Swarms, you've got multiple different agents working as a team. And you'll see what that looks like in a second, because it will begin to start loading up.

Now, bear in mind that all of these use different amounts of tokens. So if you've got a bigger context window, you're going to use more tokens on the plan that you're on, which means that you're going to be able to use it less if you're using it for bigger projects, like for example, a Swarm. If you are using the standard context window and you're just using K3 Max, not Swarm, then actually you'll use less tokens, which is even better. And you can see, it gives you updates as you go along.

So it's like, here, we're done, et cetera. You might come back and you're like, what's taking so long? So you can actually close the tab and then you can come back to it later. So you can see multiple different conversations working at the same time with Kimmy.

So the good thing about this is, yes, it might take a little bit of time to come back to us, but we can have those running in the background. We can come back whenever it's ready. You know, the agents are doing all the hard work. Really, we're at the point where you just need to give the direction.

You need to come up with the idea. And that's pretty much it. If you've got an idea in your head, I wouldn't be afraid to test it because everything that I've thrown at Kimmy, it's managed to handle it. It's managed to do really well.

So whenever you have an idea, you can go from there. And you might think, okay, what should I create with Kimmy? What can I code with it? Well, I would look at where are you spending your time?

What's wasting your time day to day? And then how can you automate that with a cool tool or a cool website or a cool app or whatever you want? So for example, if I do a lot of SEO, it's like, right, let's create a beautiful SEO app for keyword research. And we can do that with Kimmy.

We could, for example, look at our time and we're like, right, okay, I'm spending a lot of time managing my sales here. So why don't I create a tool with Kimmy that can analyze my team's calls, analyze and give feedback on how to improve based on my existing SRP. And then Kimmy can create that tool in the background, which is pretty amazing. So we've got two different tasks being created over here, as you can see.

Swarms. And what's interesting about swarms is you actually give names to each of your sub-agents and assign different tasks to each of those agents. So you see here, it says agent recruitment with swarms. It's like one agent recruited.

Here's what it's going to do. It's going to create the sub-agent for SEO keyword research. And that agent has only one task, which is coming up with the keywords for the website that we want to build over here. And whenever you are working on, whatever you're trying to automate, it will customize this to you.

And again, you don't need to think about any of this. It's basically like automatic team orchestration where your agents orchestrate them for you. So you'll probably recruit more agents in a second. It's just starting with the sub-agent, as you can see.

And on the right hand side, again, we can see Kimmy's computer. So we can see what's being done over here. And then we've got the chat over here. If you ever need to interrupt it, you can click on the stop.

And if you ever need to like give it, for example, files to deal with, excuse me, then you can upload those right here. So if you have a lot of information about your business, you want to custom train the agent on to give it more context, so it understands you, your projects, your goals. et cetera, you can add those in. We also got options for plugins and skills, which I'll come on to in a second.

And we've got goal mode, which is really interesting as well. This is a way of making Kimmy even more autonomous. So let's open up and test Kimmy.com with goal mode. I'm excited to try this.

You'll find this with every single agent out there right now. This goal mode, basically what it means is that you can give a task to Kimmy as an AI agent, and it will just keep going until the task is done. And what it actually does is there's a judge and there's an AI agent and the judge looks at the agent's work, quality controls it, checks if the criteria for your goal has been met, and if it hasn't, it will keep looping around until goal is finally met. So let's do an example here.

We're going to say, okay, create an SEO keyword research tool where I can find relevant keywords for my niche, any niche, but they must be actual real keywords and the tool must be fully customized to meet my niche. It can't be like a code-based thing. It has to be something that's fully customized where the AI actually understands what we're working on and then creates the keyword research around that. So basically what we're trying to do is build an LLM into the tool that we're creating so that we can get better outputs.

So from here, we're going to plug in the setup over here, right? So this is going to be working on goal mode, like so, and we've set that as a goal. We can also share conversations. So if you want to share this with your team, you can do it like so.

You can also cancel that and you can also copy the output as well. So just to recap, we've done quite a lot so far. We're building out a website inside the normal chat with KimmyK3Max. We are also using the agent swarms feature to have a team of agents working on a full blog website with published content.

And then over here, we have a keyword research tool that's being done and all of our agents are just working in the background. And this is a beautiful thing about Kimmy. It's like, you can give it tasks. It works in the background.

It's all autonomous. And it's so different to using, for example, ChatGBT. Like if you're using ChatGBT and you're just asking it to do stuff over here. It's slow and it also requires you to sort of babysit it.

Whereas for example, with Kimmy, it's a lot more autonomous. It just works in the background. It goes off and does stuff. And then, you know, you can check that later and see what's being done.

And also when you're running the goal mode, you'll see how long it's been running here. And like KimmyK3 is so powerful that it can run autonomously for days and hours. If you just leave it like this, it's pretty wild. Now there's some basics that we need to go over as well with Kimmy.

So if you wanted to generate slide presentations, you'll actually see some examples here. So for example, we can select freestyle or we can go with some of the templates for slides. So if you need to create a presentation, this is a fantastic way to do it. Lots of cool visual slides here.

I will say KimmyK3 is very, very good at the visual effects and very good at design. It's actually number one in the front end code leaderboard. So it's great for presentations. So we can also say, okay, create a beautiful slide deck about the latest updates from KimmyK3.

And so we can create a presentation about KimmyK3 just as a little example over here. And then something to note as well is that Kimmy can also generate images. So we can say, create an image of a cat eating biscuits. So if you need to generate images, you can do it inside the chat here if you want.

And also there is a limit to the number of chats that you can use as well at the same time. So right now we have one, two, three, four different chats going at the same time. That seems to be the limit on the current plan that I'm on. I'm on the Allegretto plan and depending on which plan you're on, you can switch between them and test different things.

So just so I can keep showing you, I'm actually going to upgrade as you can see here. So these are the plans and you can choose between them. I think there's also a free version of Kimmy as well. So they do have a free plan, which is pretty generous as well.

There is a free tier. I'm just going to upgrade so I can show you more cool stuff at the same time. So if we take a look here, we've got Kimmy now ready to go and we can set up an image. So if we say, okay, create a image of a dragon in the sky.

The prompts as well can be really simple. The AI is smart enough to understand what you're working on. So you don't need like any fancy prompting techniques or nothing. We're at that point.

Honestly, at this point as well, here's the interesting thing. I would say we are at the point now where your agents just run on loops, which means you don't need to prompt anymore. So this is a recurring theme here. So I don't need to prompt the goal mode for creating the keyword research tool.

Instead, it just runs on the loop in a background until it's finally done. And that's the way that I would see working with agents. You shouldn't be looking for the best prompts or asking what the best prompts are in a file app. Just focus on giving the right loop, designing the right loop so that the agent can meet that goal based on you designing and architecting a beautiful loop for it to follow.

So this is now asking, okay, how long should the slides be? What visual style do you prefer? Then it's going to start running through. Now you'll see here that it has all these skills for understanding how to complete certain tasks.

So it has, for example, slide categories as PowerPoint MD. These are markdown files, .md files. And so it basically shows the agent how to do certain things, certain workflows. You can actually create your own skills as well.

So if you go to a new tab here and then we're going to click on skills and we can add our own skills. We can create new skills with Kimmy. We can upload them and we can also import skills as well. As you can see right here, there's all sorts of cool skills that you can actually get straight from the chat.

So if we go on the skills section, you'll see, for example, we've got a WhatsApp integration skill, like all sorts of interesting stuff, pitch, deck, creator skill, ad creative skill, like lots of cool stuff that you can have a look. Even browser. So you can get Kimmy to browse the web for you using that browser skill. And you can import that in.

It's all ready to go and preloaded into Kimmy as well. You can also click on manage skills over here and you can have a look and see, okay, what skills are there out there that I want to import in. So we can click on use, or we can click on add and we can actually test those. And if we like them, then we can add them.

And if you want to search through the skills marketplace, you can do that over here. So if you're like really into productivity, well, there's tons of skills for that. If you're really into marketing, loads of marketing skills over here. And the way that I would use them is like use a skill, but then tweak it based on once you've tested it, right?

So if we, for example, test out the copy editing skill and we're like, you know what, that's great, but here's how I'd like to improve it. Then you can easily update the skill whenever you want to as well, which is really good. Now you can also see the website is almost ready to go. So it says the entire site is verified and looks polished end to end.

So Kimmy can actually look at its own website and see what it's doing so far, which is great. And then we've got the full page workflow over here. And then it's going to open up a preview once that's completed. So it's going to take a look and we'll be able to preview it.

Plus we can get the code. So you'll see the code section here. If you actually wanted to, for example, share this with someone, you can click on the message box. You can also publish this to the web.

So if you create an awesome website, you can publish it, right? So we can hit publish there and now people can access the website that we've created. So we stored on a sub domain of Kimmy, which is .kimmy.page. And then we've got the name of the domain over here, which has taken a little while to preview.

In the meantime, you can see our team of agents, our sub agents, our team of agents, our sub agents are now fully being built out. So we have multiple different agents. So it's now recruiting four different agents at the same time. We've got, for example, Barfez, who's writing posts.

We've got Tej, who's writing more posts. We also have Bauman and we've got a pro designer, Feynman, who is creating the blogs as well. So these agents are all working in the background. You can see their progress over here too.

You can also stop them at any time if you prefer, and you can hide them as well. So if you want to see what agents built are, you can click on that. If you want to hide them, you can go over here and you can click through to each of these sub agents and see what's being done. And again, like this is awesome for massive tasks that require a lot of work.

So now we've got the website published. If we click on this, this is the publicly viewable website. And I mean, look at that. Like it looks beautiful.

We didn't have to do anything. It wrote the content, it designed it. It customized it to our search intent, which was SEO. And it even created all these custom graphics and put it all together in a beautiful landing page.

Now, if you actually click on the navigation bar as well, it's not like a full website, that's what we're actually building with the Kimmy Swarm, but like all the dropdown boxes here work beautifully as well. We've published it to a custom domain. It also has a free SEO audit link at the top as well. Looking fantastic.

I really like the fact that it's got these scrolling panels as well. One thing that we'll say is you always want to just quality control it, have a look at it. Obviously in a situation like this, if I was really building it out for my agency, I would give the details to my agency, my branding, the link that we want to include for the free SEO audit. But for a first attempt, not bad at all.

Now, if you wanted to, for example, publish this to a website, you could actually give Kimmy K3 Max access to plugins on the web to go off and do stuff for you. So if we click on manage plugins over here, you can see that we've got all these different types of plugins that we can add into Kimmy. So for example, we have image generation, we have academic data, audio generation. So it can actually create audio for us as well, which is unbelievable when you think about it.

Then if we click on all, we can browse all of the available plugins. And one thing that you'll see over here is super basic. Now Superbase is really good, you can actually install that and when you install it that means you can add a database so that people can log in to your website or app or game whatever you've created. You can also click and add this to github so you could publish it on github and make it open source for the website itself.

You can also generate videos, images and we have the option to use this on Cloudflare so you could host the website itself on Cloudflare when using this and also this is great for security as well. Now if you want to install any of those plugins you can just click on install and you're ready to go. So if we click on install on that it's then going to ask us to log in and accept permissions for this and then we can grant access to the Cloudflare API for us. We've still got the slide deck being created over here and we have the image generated for the dragon in the sky.

So if we take a look at this, it takes a little bit to load, you can download it like so if you prefer, that might be a bit faster for you. You can see like the images it creates are really beautiful as well and this is all included in the plan. It can really do so much, I think it's very underrated as a tool and bear in mind like this is already like one of the best AI models in the world, K through itself, let alone the fact that it can generate audio, it can generate images, it can generate video and these are things that for example chat GPT, it can generate images but it can't generate video and then you've got for example Claude and that can't generate video or audio or images but KimiK3 can do all of them and so you see how number one it can create potentially better outputs than Fable 5 and GPT 5.6 but at the same time it can also do so much more than that. Which brings me on to the next part of the process.

So what we have over here is we have OpenCLAW working directly with KimiK3. So if you're wondering what OpenCLAW is, this is an open source project which is an AI agent. Now the problem with OpenCLAW is that as fantastic as it is, a lot of people struggle to set it up. Most people don't know how to use GitHub and if you're non-technical, don't worry because what you're about to see totally solves that problem.

So what you can see over here is we have KimiK3 and we can create a new agent just by clicking on add CLAW. We can deploy it to our computer with the desktop app, we can have it on our Android phone as well or we can actually create it on a cloud server. So we can create a new CLAW like so, we can give it a name like for example Julian. We can then switch between the type of voice that we want to plug in here, then we can click on next and then we can type in what we want to do with this.

So for example like content creation, workplace productivity, those seem like two things I would like to focus on and literally in like 60 seconds we'll have a OpenCLAW deployed in the cloud that we can use from our phone and we can easily set it up like so. And so it's almost like there's really nothing KimiK3 can't do. Now if you're wondering, okay what is the desktop app? That sounds fun.

So if you've ever come across CloudCowork or Codex for example this is basically Kimi's answer to that. It's a desktop app that you can operate directly here as you can see. You can update it as well and then you can see that it knows which plan we're on and we can use it in exactly the same way. But the difference is with Kimi on the desktop app is that it's coding locally.

It can access things locally and that also means that it can run 24-7 locally on our computer. So we can select what permissions it has here and bear in mind because it's working locally it's going to ask you which permissions you want to give it. Because obviously if it's got access to your local files, to your local folders, it can make some changes that could for example delete a file or overwrite a file and so that's why it asks for permissions here. You can switch between asking permissions before taking each action or if you're an advanced user you can just go with full access.

So we're gonna allow full permissions right here and then we can use the same, it feels and looks exactly like the same system. I'm actually going to switch over to dark mode here and we can also allow access in the browser too. Then we can change our theme over to dark mode. That's a bit better isn't it?

And we've got all these different presets that we can use, we have all these different things. Something else you'll see here is that we can add memory instructions. So what this means is that you can give context to your AI agent so they understand everything about you and I'll come on to a system that I prefer to use for that in a second but you can basically add different memories to Kimi work and then it will understand everything about you, who you are, you can add different memory instructions there too. You can also use it inside the chat and say save this, do that etc and then we can also connect this to our browser.

So we can connect this like so and then it can use our browser and that's basically it for the settings. So we have our dashboard over here with some examples of what we can build and this is pretty cool because we can create a custom dashboard inside Kimi that we can have a look at whenever we want to. So for example we've got this option here and we've got all these different widgets that we can come back to in our dashboard. So if we have a look we've got for example our cover hero image, we can resize these, we can change these, we can edit it like so, we can decide what we want to show on the information and then we can also pin this and we can edit the whole setup too.

Now if we go back to your new task, we can get it to work inside a project. So if we click on like new existing folder and like just it's a totally random example, we'll say like neon snake game, bear in mind like hopefully if you're watching this you're building something more advanced but I'm doing this for a demo for a course so that's why I'm just going with something quite basic and then we'll say from here create a beautiful neon snake game in HTML and CSS, let's launch it for me and it will start working on that and then we have the project and we have the chat. If you're familiar with codecs this looks and feels exactly like codecs. Now what you're also going to see here is you have the chats as well.

So we have the project and we have chats. Chats are just like back and forth and then projects are things that you come back to because they're more long-term projects and this again is inside the desktop app so this is installed locally and then we can go back to our browser and we can use it like so. You can see it's connecting the services for OpenCLoud which we'll come on to in a second. So now we're beginning to have our website completed and you can see this will take quite a long time so this is what I mean by swarms working autonomously for hours.

You can see that it's on task progress one out of four which means it's going to be really in-depth. We have the SEO keyword research tool over here though so let's have a look where it's up to like so. It says done your keyword research tool is ready and you can see that it's kind of got confused here so it's definitely not perfect and I think that's good for you to see because then you can see okay this is where it messes up sometimes. So I'm going to say make sure you turn this into an actual web app that I can use for keyword research so that we can give a front end to it as well.

We don't just want to create like a basic markdown skill file for this. Over here we have the slide deck still being created and if you want to generate a video as well we can just test it out now. So I'm going to say create a video of a dragon fly through the sky and we'll just edit that prompt quickly for you there. By the way when you're prompting this like don't worry about grammar or anything like that you can just go with the first.

Usually because it's an LLM it's designed to understand language which means it can understand what you're trying to say at any point. Then if we open up a new tab here we have the open call that we were talking about a second ago. So this is the AI agent that we can actually connect to our phone. So if we click on settings over here we can edit the name, we can open up the terminal, we can restart it if it gets bogged down and then we can also see the usage.

We can reset the instance and then we have the backup and you can actually restore to a last backup. So if you create a backup over here then you can restore it and if we scroll up we can now connect this to a channel. So we can actually connect this to any of these channels as you can see right here and it will basically connect to those channels so that we can use this on our phone whenever we want to as well which is pretty nice. Now inside the chat we could say something like schedule a task every day at 5 p.m.

and then go off and do that. So if we say for example scheduling the task of creating a new Facebook ad for an SEO agency in vertical format as an image and do that every day at 5 p.m. give me like three different options and make sure it's optimized to say free SEO strategy session. For the actual copy inside the ad make sure that you also write in a Dan Kennedy style of copywriting and additionally if you could give me one video too that would be great.

So now we can schedule in these tasks to run energetically in the cloud whenever we want to and we can come back to that like so and we can also connect this. Now when you're doing this as well you'll see that the preset channels over here when you're connecting the channel these are all kind of like Chinese channels so WeChat you know QQ these are not probably channels that a lot of people watching this would use. So if you actually want to connect it to telegram you can actually ask it inside the chat okay tell me how to connect you to telegram it will ask for your access token to telegram or to whatsapp or whatever and you can connect it directly there you don't need to use the preset channels you can ask inside the chat to do this stuff for you. So it's beginning to do that in the background over here and then now we have all of these agents running like so.

Now if you go back to Kimi inside our local setup here. We can also create a new task now, and I'll show you another example of how you can use it. So we'll create a new folder here, and I'm going to click on start from scratch, and we'll call this Hermes Agent. So if we type in that, and then what I'm going to do is I'm actually going to take a GitHub account, which is this one for Hermes Agent.

This is another open source AI agent that's very popular right now. And what I'm going to do over here is I'm going to plug that into Kimi, and I'm going to say, set up Hermes Agent and make sure that Kimi coding is plugged in as OAuth into it so that I can use Kimi's coding plan with K3 inside Hermes Agent, and then just configure it for me so that I can use Hermes Agent. So we can go from there and hit enter, and now we have a lot of different projects being built. Right, we have the snake game, we have Hermes Agent that's going to be configured with the API, well, the model and the OAuth for Kimi.

One thing to note as well is like, when you're using AI agents, if you're using Claude, for example, you have to pay for the API. But if you want to plug in Kimi into an agent, like for example, Hermes or OpenClaude, you don't need to pay for the API, you just use your coding plan and then you can log in with OAuth. So it's now going to start looking at this particular page and set up the workspace. And also the good thing about this is because we've got it set up as a project, we can come back to it later, we can use it later, we can add new things to Hermes, but we've got that full conversation to come back to that's fully trained with context on exactly what we do and how we work.

Now, if you go back inside here, what you can actually see is that it's not responding to us inside the OpenClaude chat. So I'm just going to say, did you set up the schedule task? And one other thing that's pretty important to note is that we have Kimi code. So there's so many different variations of using Kimi.

Honestly, I think for you, like out of everything that we're seeing here, I think if you're a beginner, then you're just going to use chat. If you want something that's a bit more advanced, you've got bigger projects to work on, then I would use something like Kimi Swarms. And if you want something agentic that you can handle from your phone, that you can operate from your phone, then I would use OpenClaude with Kimi Claude. And if you are, for example, working on as a coder, and you're building out, for example, something bigger like a mobile app or a project that you're going to develop and work on every single day, and I'll show you an example of that in a second, then I would use something like Kimi with the desktop app.

So depending on where you're at and what you're trying to do, you're going to use it in different ways. And that's why they've created so many different use cases. So if you're confused, just to recap, you're going to use the chat with Hermes for, sorry, the chat with Kimi for normal tasks, back and forth for your beginner. You'll use Swarms for big tasks.

You'll use Hermes desktop for local projects you're working on day to day. And then you will use the Kimi Claude setup for agenda tasks on your phone. And this is a big advantage, by the way, because like, for example, using Claude code on your phone, total nightmare, total nightmare. It's not particularly good at all.

Whereas for example, open Claude on your phone, super easy, straightforward, good to go. Now, one final thing to note is that you have a CLI. This is kind of like Claude. So if you ever use Claude code in the terminal like this, so if you go to the terminal here and we type in Claude, this will open up a terminal-based CLI for Claude.

We can do the same thing with Kimi. How would you do that? You just copy this command from the Kimi code page. So if we copy that, and then we go over to the terminal, we're going to paste in that command like so.

And what that's going to do is now set up and make sure it updates to the latest version of Kimi code. If you don't already have Kimi code installed, then it will install it for you directly. As you can see here, it's updating. So that might take a few minutes, but basically what that's going to do is install the terminal command line set up for Kimi code.

We also have this available inside IDE. So what this means essentially is if you have a tool that you prefer to code with, like some people watching this might prefer Visual Studio Code, then you can plug in Kimi IDE into this as well. For me personally, I don't worry about that. I don't really use the terminal so much, but I will show you a way in a second that I prefer to use Kimi CLI and Kimi code, which is even better in my opinion, but we'll come on to that in a second.

So if we have a look over here, Homeization is getting set up nicely. We actually have the snake game ready to go, so we can preview that. And if we actually click on Preview, you can see that it's not working. So what we're going to do over here is check the local host and you can see it's not working.

So we'll go back to Kimi and we'll just say, can you make sure this is up and running? Then open it up in Chrome for me. And this is something to note, like as smart as these agents are, there's going to be bugs, there's going to be times where they break. If that happens, don't get frustrated, just realize like you're super early to AI and this is part of the journey.

Believe me, 12 months ago, it was even worse. It's only getting better to be honest. We also have the presentation that we can view over here. So we were working on that presentation about Kimi K3 and look at that, like it's created a full presentation, looks beautiful, it's actually researched all of the benchmarks.

So it's done a lot of deep research. It's talked about Kimi Delta retention, which is pretty interesting. That's part of the report card for the Kimi model and how it works, et cetera. And then it talks about all the ways that Kimi runs.

So for example, Kimi.com, Kimi work, Kimi code, Kimi app. And it's created a beautiful presentation as you can see right here. And what we can do is we can share that with people, but we can also play it. So we can preview it in full screen and take a look and see what it looks like.

But it is looking really, really nice. And we can export this. So if we click on export, we can decide whether we want it as images or slides, whether we want to embed the fonts, how we want the slide transition to be, and then we can download it, or we can share it, or we can actually export it to Google Slides. So if we click on export to Google Slides, then we'll just log in and we can connect Kimi to Google Slides.

You can create the presentation. We just have to allow pop-ups and now it's gonna open up the presentation. And how easy was that, right? It's created a beautiful presentation.

If you have something coming up, maybe a public speech or something like this, you can use this for the first draft. Now you can see by the way it writes that it doesn't have a lot of personality. So it's just kind of giving us the facts. Some interesting things to note about Kimi is like it's a 2.8 trillion parameter model, which actually makes it the biggest and largest open source model in its class.

But as you can see right here, you can easily automate this whole setup. Beautiful. Works first time around. So it's smooth in a lot of cases.

Now you can actually see, if we go into Kimi Claw Excel, this is not working for us right now. And I probably think that's because Kimi K3 has just launched and probably the servers are a bit overwhelmed. So just something to note there is like, it's not that smooth on the OpenClaw setup, but we can install Hermes locally as well. And I would expect that to be much smoother in the future.

It's just because we're in that period between Kimi K3 launching and the whole world using it. And that's why it's happened. So if you want to see an example of it actually working, you can see an example right here. And so we have OpenClaw with a previous setup that we created, and it responds to us inside the chat.

It goes back and forth for us, right? Now, if we go back to the terminal and we want to use Kimi for creating stuff, we can type in Kimi here, and that will launch the CLI once we've installed it. So you just do that one copy and paste command to install it, and then you launch Kimi, not Kipi. And then from here, we'll just carry on, and you can see that we're using this, like so.

Now, what you also see is that the model inside Kimi is the default as K2.7, right? So K2.7. However, if you want to switch to Kimi K3, all you do is you type in forward slash model, like so, and then it will give you the option to switch over to K3, and you can switch over to K3, like so. That's how you can switch between them.

Now, if we have a look over here, we've got the Snake game ready to go. It looks pretty easy, looks pretty nice, and it's working as well. We literally just created that with one single prompt, which is awesome. And just testing out, works really nicely, actually, as well.

So you can see how you can build mini apps just with one prompt. And then if it doesn't open or you have any problems with the preview, just say, can you make sure you get this up and running, and then open it up in Chrome. I do think it's important for you to see all the bugs that come alongside this as well, because that's going to happen. It's just natural, it's part of the game.

Now, what we also have over here is something called Kimi Web Bridge. Kimi Web Bridge is a way to control your browser with Kimi Agent. So we can install the browser extension, and only do this if you're comfortable with AI controlling your browser, but we can add the extension, as you can see right here. And then this means that with Kimi Desktop, once we've set this up, so let's just make sure that we're connected here.

Yes, we are. Then we can get Kimi to control our browser. So if we click on the new chat, and then we say, open up Google inside my browser, we'll hit enter like so, and that should use the browser extension to open up. So we've got Kimi Desktop over here, and that should be connected to our browser, so we can use browser automation.

You can see here that it's using the skill for browser use, Web Bridge, and it's loaded that skill, and then it's going to use this to reply and start working for us. Now, you will see here that it's a little bit slow to reply. So it's setting up in the background, and we can have that running in the background. Let's see what we're up to with Homies Agent.

So it's setting that up as well. Neon Snake game completed pretty nicely, and we just have to enable Kimi Web Bridge as well, and then we'll pin that as well, just so I can see and monitor it. So that should be working. Let's say test again.

So you can see here that it's now working. And what you actually see when the browser use part is working is that you'll get a selection like this, right? So there's like this orange underline and then the tab that's being controlled by the AI is right here. So you can see the Kiwi is basically controlling that.

And if we want to control our browser even more, we can ask it to do that. Now you might say, why would you use browser use? Well, basically you can post and you can get the AI agent to, for example, like automate your social media or that sort of thing. So you can give it a command over here and then it can post into your group like so, which is pretty awesome.

So we'll close that now. If we want to uninstall the web bridge plugin, we can just click remove and we can remove that. And now it can't control our browser anymore. Now let's come back to that video we were creating before.

And look at this, this is really interesting. So it created kind of like a cartoon style video, actually. That wasn't what I was aiming for, but it does look very nice. Like it's quite interesting that you can generate videos like that.

And if we come back to Kimmy Claw, which again was a bit buggy. So I just want to be a hundred percent honest with you. I can't control that, but I will show you what it's done here. So it's actually set up the schedule task.

It's got the job ID, the schedule. We can cancel those at any time, but basically that will loop around on a 24 hour basis. It's going to create images, it's going to create videos for us, specifically for our Facebook ads campaign, which is pretty awesome in itself. Now you can also create scheduled tasks inside Kimmy desktop and also inside Chrome.

Here's the difference. Kimmy.com is in the cloud. So if you're not having your computer open 24 seven, if you are running scheduled tasks whilst you're sleeping, if for example, Kimmy desktop, it's not going to be open for you all the time, then you're better off creating a scheduled task inside the chat version, inside Kimmy.com instead. And you can choose between creating manually or creating by the chat.

So if you're not really sure what you want, but you know you need something, you can just create a scheduled task like this. You might say, why would you schedule a task in? Because for example, you can just have this running 24 seven and then you check it when you need to. So it's another loop where you don't have to prompt the AI agent to do stuff every single day for you.

If you look at all the tasks that you do, you've probably got a certain amount that are basically just repeating tasks running every single day. So for example, we could say to Kimmy here, can you create a scheduled task where you research the latest AI automation news? And then you give me a bunch of content ideas based on what's just dropped over the last 24 to 48 hours. Break it down into like a beautiful sort of bullet point list as well with sources so that I can understand exactly what I can create content around.

And so we can create the scheduled tasks like so, we'll hit enter, and then that'll guide you through the process. You can create this manually. So if you don't want to run through the wizard, you can go back to the scheduled tasks here, just click on create manually, and then plug in the details. So you enter the name, the requirement of what you want it to do, how often you want it to do.

So it could be daily, it could be weekly, it could be monthly, it could be one-off, and then also the time, and also when it's going to expire. So for example, if you just need this automation to run like every week, or for example, for the next seven days, then you can run it like so. So it's really good for automation. If you come back to the AI automation ideas chat here, it's going to guide you through the process.

It's like, okay, how often do you want it? When do you want it? I said 9 a.m., right? We'll hit enter, and now it's just going to start running that directly with us.

Now let's talk about some more advanced stuff here, because we've talked about pretty much the basics of everything, and how you can, for example, use Kimmy Claw, Kimmy Code, Kimmy Work, Scheduled Tasks, plugins, Kimmy Desktop, the CLI, how you can generate images, videos, slide decks, websites, apps, tools, games, whatever you want. I've shown you pretty much how to automate anything. There's really not much to it. Like it's very simple once you get your head around.

Now what I want to show you is a system where you can use Kimmy Code in an even more powerful way. And I think this is one of my favorite ways to do it, because for example, if I have to like click through to Kimmy, and then I have to click through to Claude, and then I've got Hermes agent running over here, super messy, time consuming, a lot of button clicking for absolutely no reason. So what you can do instead is you can build out an agent operating system like this, where you have all of your agents, including Kimmy Code plugged into this system. And again, the beautiful thing about all of this is that you don't need to pay for the API.

When you get a subscription for a tool, like for example, ChatGPD, or Claude, or Kimmy, well, you can use the CLI for no extra cost. It's just included inside your subscription. So as soon as Kimmy Code 3 dropped, we actually plugged it into our agent operating system, and we can use it right here. We can switch between quality, fast, and no think.

We also have the same system inside, for example, Claude and Hermes as well. Now also, we have a setup with Hermes and Kimmy Code 3 over here. And from what I've tested, it is as you've probably seen today, like when it's on the thinking and reasoning model, it's not that fast to run, but if you're running it on a scheduled task, which I'll show you in a second, then that doesn't even matter because you're not going to be there to be waiting for it. So this is Kimmy Code 3 setup with Hermes agent, and it's plugged into our agent operating system, and it can run tool calls pretty well.

So for example, we said before, forward slash learn, then we gave it access to a tutorial of ours, and it said, learned, here's what I found from the guide. It broke down my guide on how to do long horizon coding with Kimmy. And then what I actually did is create a skill and save that for later. So this is Hermes agent with Kimmy Code 3 plugged in.

And again, Kimmy is using the CLI. And then for the agent, that's just a free open source agent. So the brain and the hands don't cost us anything extra, but we can use it in a better system that's way more organized. I'll show you a couple of other examples.

So over here, what we have is Hermes Apollo, and this is a voice activated AI agent that runs in real time, and it listens to everything that I'm saying. So you can see that it's actually typing out the sentence that I'm saying right here. So if we go into Hermes Apollo, and we're like, open up Google for me, please, sir. And it just opens it up as you can see.

So you can use browser use. We can say, open up Obsidian for me, please. And it just opens up Obsidian on our computer. So you can use computer use and browser use, and it's ready to go.

By the way, this is our memory system for giving context to our agents, which I'll come on to in a second. So these are some more interesting workflows that are more advanced, but much more powerful, and help you get like 10 times more power for AI agents. Like if you're just using Kimmy.com, it's cool. It can create amazing stuff.

Don't get me wrong. But why not have that sync to all your AI agents? Why not have that inside one system that's ready to go? Why not have a voice activated AI agent that basically operates in the background whilst you're talking to it and doing something else?

We've got, for example, Hermes Oracle as well here. And this can use Kimmy K3 to pull in the latest news from the web, kind of like we were automating before, but it does it in like this beautiful UI. And then also it gives us ideas on what are the most trending topics in your industry, organize it by what's trending right now. And then additionally, we can draft social media content or we can publish content with SEO directly to our website using this system as well.

And this runs on a 24 seven loop. So we can see, for example, Thursday, Wednesday, Friday, Tuesday, Monday, we can see all the previous versions of this system. And then what's really cool about this, if I want to draft social media content about this particular topic, then I can just click on draft right there. If we want to check out the blog posts that we previously created, so I actually ran this workflow this morning, you can see the Kimmy K3 topped front-end code arena.

So what did we do? We created three blog posts talking about that topic so that we can rank it on Google. So for example, you can see here that we actually created this article and it is already indexed on Google. So this is an article we created with Kimmy using Hermes Agent.

It created an article on a website, published it directly there. And we literally just give the system our WordPress login details and the API access token. And then it goes from there, right? Pretty simple and easy to use, but you can see within seven hours, it's already indexed in our Google.

Not only that, but for that article that we created, so for that keyword, we can see that we are already ranking on the first page of Google for that particular keyword. So this system works. It's pretty good for creating and ranking content quickly. It's all fully optimized and personalized to our brand.

And you can get the same system. We have it inside the AI Profitable Boardroom if you want to get it. Link in the comments description or go to theaiprofitableboardroom.com. That's my AI automation community that helps you save time and grow and scale with AI automation.

We also have this workflow just to give you inspiration, really. I want to show you what's possible so that you're not just like stuck in that bubble of just using Kimmy.com or using Kimmy Desktop because there's so many more possibilities for this. So what we have over here is Hermes Astros. This is another custom workflow that we created.

And you can see, for example, that it can actually monitor our competitors, look at ideas for content, and then give us new unique angles for that topic. And also we can plug that into our video agent, ICO content agent, or notebook LM to generate infographics or to generate videos or to generate SEO content for that particular topic. And it's all inside one ecosystem. So you see how, number one, you can move way faster because you have all these workflows ready to go.

I don't have to go into Hermes Desktop and then open up OpenCLR and then hope OpenCLR replies to me. It's like, no, no, no. We've got all of this running on a loop, 24-7, scheduled in, very tight system, and it just runs whenever we need it. The same, for example, with the studio.

So we've got Grok and Minimax plugged into this, but you could use Kimmy K3 as well because it can generate images, videos, and voice. And so we can switch between these tabs and we can generate image or video or voice whenever we want to using Hermes Agent. It's basically like a SaaS tool for generating this stuff, but it's all inside one system that all links together. So for example, if we create something like an image over here inside Hermes, we can actually go up to Claude and say, create a beautiful website from this.

Or we could create a video over here and we could go up to Claude and say, hey, create an app or a game with this video that we've created. We've also got, for example, an hour each agent. This is really good if you want to use KimiK3 for lead generation. All you need to do is plug in a Hunter API and then you can generate leads on the spot using this system.

So we can type in whatever type of leads we're trying to generate or find. So if you're doing any sort of outreach for any sort of email outreach and you need leads, which pretty much every single business does, we've got this system where you can type in whatever type of leads you want to find. It then pulls up a leads list as you can see. And then from there, we can actually automate drafting the email campaign and that is fully personalized to the leads because all of the leads get enriched by the agent that we're using inside this.

And then from there, we can actually send emails. We've got the full flow that we can preview over there. We can manage our campaigns. We can preview them, pause them.

And the beautiful thing about Hermes Agent as well is that with the power of KimiK3, it can also manage our email, send our email inbox and handle everything for us. And then we get the stats over here. So it's like a lead generation tool, but it's plugged into a Hermes Agent system where we can come back to that anytime and it can all be powered by KimiK3 and the APIs you want to connect with it. The same, for example, if you want to link KimiK3 to external different tools.

So for example, we've got Blender, which is like a 3D model system and Blender itself is super, super hard to use. Like it can take hours just to learn the first tutorial with Blender. Most people give up halfway. However, because we've got Blender MCP enabled with Hermes Agent, we can get Hermes Agent with KimiK3 to create amazing 3D models.

And for example, like product shoots using Blender. Blender is free. Hermes is free. KimiK3 is included on our plan.

And then we're using a powerful modeling system that normally you'd have to have a really, really technical person to create that for you. And then it's working alongside all these different agents here. You could even, for example, get Cloud Code. If you're familiar with Cloud Code and you really like that agent harness, you like coding with that, what you can actually do is you can plug in KimiK3.

So we've got GLM 5.2 and GPT 5.6 Sol plugged into this, but you could also plug in KimiK3, which is the most popular app being used with KimiK3 right now on OpenRouter in the whole world. So it's very popular to do this, even though it's only just dropped 24 hours ago. And so with that, you can have Kimi inside Cloud Code and then it can create stuff and you can actually preview it inside your workspace right here. So all these different things that we've created, we've called code as a harness, but we can plug in a different brain like Kimi into that as well.

And then also you can, for example, like get Kimi to orchestrate all these different open source projects for you. So for example, OpenMontage, really popular film, like AI film creator, as you can see here. So you can create these beautiful films like you can see that we can preview and play. And when we're doing this, essentially we can get KimiK3 to create this for us.

We just type in the type of video we want to build and then we can preview over here, which is pretty cool. Finally, we have a memory system. And what you're going to find with Kimi is that straight out of the box, it doesn't know anything about you. And if you've been using, for example, Cloud or ChatGPT, a big problem is that when you're creating content, it doesn't understand you, it doesn't understand your goals, it doesn't understand what you're about.

So how can you link it to a system where it actually understands everything about you? This is why we use something like Obsidian. And Obsidian is just like a way of storing in very simple, very simple, like Markdown files, everything about you. And whatever agent you use right now, if you want to switch over to Kimi, or if you want to try them all, but give them instant context, you can actually download for free something like Obsidian.

You can create a knowledge base like this, and you can see, for example, every project that I work on, every tool, every app, et cetera, is all plugged into this one system. And the great thing about that is that we can understand how our agents gather information about us, how they document it, all the information they have about us. And also, this can be created by your AI agent. So literally, to create this whole system, which I actually call my memory galaxy, where all of these different nodes are memories that interlink together.

When we're doing this, essentially, we can get something like Hermes to just document once a day into Obsidian everything it knows about us, plug it into the memory system, and then we've got context across all our agents. So Kimi wants to know about us, or if we need personalized recommendations from Kimi, we can just go straight to Kimi and say, hey, knowing what you know about me, based on my Obsidian memory, make sure that you give me personalized recommendations. And we can do the same for any agent, basically. Final thing that we can do is we can actually build teams of AI agents with Kimi and Hermes.

So what you can see here is this is a full AI agent company. And again, I just want to inspire you with some crazy workflows that you've probably never seen before. So we have Paperclip, and this is another open source project that you can plug into your agent operating system. And what you can see here is that we have a full organization map of AI agents working together.

So what we can have, for example, is we can have a chief technology officer, and then that's managing a bunch of agents. Underneath it, we have a CEO that runs the whole company, and they can all be powered by Kimi K3. So that's something else that you can get as well. This is all inside the AI Profitable Boarding, by the way, if you want to get my systems for doing this.

You could create it yourself. It would take a lot of time. Personally, for me, I code this out for about three to four hours a day, just improving it, tweaking it, making it better and better over time. Taking the latest updates, for example, like Kimi K3, and plug it into the system.

And if you want to get my full system for that, you can get that inside the AI Profitable Boarding. Link in the comments description, or just go to the AIProfitableBoarding.com. This is my AI automation community that helps you learn, save time, and grow with AI automation. So if you liked this training, we actually have loads of step-by-step courses and guides, like you can see right here for all the new agents that are coming out.

Inside the classroom, you get all of our best trainings. If you're a complete beginner, you can go from beginner to expert over here. And if you want our new daily updates, you can get that inside this section. So you can get the full AgentOS.

We update it daily. You can get a full video tutorial and guide on how to set it up. Then you can also get a zip file to install it. Then every time something new and useful drops, we add it as a tutorial inside the new daily update section.

Inside the community, you can ask questions. There's 3,900 members inside there and 119 people online, which means there's always people online 24-7 to help and support you whenever you need it. And also you can ask questions, and I personally answer them every single day. Plus, I create a video tutorial answering each of those questions as well.

And then inside the calendar, you can drop a weekly coaching course, get help and support in real time. We have four weekly coaching courses where you can create a screen, meet other members, meet other cool people, build in with KimmyK3 and other agents like Hermes. And also inside the map, you can meet people in your local area who are building with AI agents like you. So feel free to get it all.

Link in the comments description or just go to the AIProfitBond.com. To recap on this course, you can use Kimmy to build out projects like the website we showed you earlier, just inside Kimmy.com. If you want to build a swarm of agents, you can do it like so. It does take several hours, so you can see 25% of the way through.

I'm actually going to stop it there, but you can see how powerful it is. And that can literally run for hours and go very, very in-depth on everything that you do. And then you can actually preview it. So if you ever stop it, you can come back to it later and you can see everything that's been done.

And all these posts have actually been created, which is pretty cool. And then also you can see over here, we have the keyword research tool that is powered with AI. So we can plug in all these details and then it will actually generate the keywords for us. And we've basically just generated like a mini SaaS app using this system right here.

Over here, we have the slide and presentation creator. So it can create beautiful slides. It can research the web. It can do deep research and it can also give us a beautifully designed presentation, which we can then send over to Google Slides to present and use later.

So anything that you need to present on, that's really useful. Over here, we have the schedule task system, where you can schedule tasks for later. So if you've got something recurring that you do daily, you can do it inside this system. And then we have, for example, the skills and the plugins.

So anytime you run a workflow that's actually useful and you might use in the future, you can save that as a skill. So if we, for example, create SEO content and I've spent ages tweaking the agent to understand how I create content, I don't want to rebrief it in the future. And so what I can do is just save it as a skill and document it like so. Then we have Kimi WebBridge, which is awesome for computer use and that can be linked to the Kimi desktop.

Inside Kimi desktop, you can have multiple projects coding locally. The great thing about that is all your files are stored locally and it can use computer use as well. And we can also have scheduled tasks and plugins linked over here as well. And then all these different projects that we create, we can come back to later.

So for example, over here, we have set up Homie's agent successfully, and then we can just run it and get it used as we can see here. And then finally, we have Kimi Claw. Kimi Claw allows you to use Open Claw in the cloud. You can come back to it whenever you want to and it can link to your phone and you can run automations like that, as you can see here.

So this is really cool. And you can have multiple different AI agents, but basically it's like a very simplified way to run Open Claw in the cloud. And finally, we have the agent operating system, which I honestly think is by far the most powerful and time efficient way to use this. And bear in mind that you can build any custom workflows you want into the system.

You just ask Kimi to build them. You could go into Kimi desktop and say, hey, create a new custom workflow for generating videos inside my agent OS. And you can actually use our zip file and install it from the AI Profitable Boardroom to get that up and running very quickly. In fact, we have over 200 pages of wins, testimonials, people learning and growing from the AI Profitable Boardroom.

So I know like if I'm non-technical, I'm not a coder or a programmer and I can use this. If you can use it too, then we can all do this. And also the great thing about that is you just meet fantastic people. You make great friends.

You meet people who are in similar industry to you. And just having those connections alone is awesome. But being in a positive environment where you can share, you can learn, you can grow together. I think that's absolutely amazing.

I've met some great people from the Air Puffer Ballroom as well. So hope to see you inside there. Thanks for watching. Link in the comments description to the Air Puffer Ballroom.

That is, Kimmy, our full one hour course. Hope you enjoyed it and let me know what you want me to cover in the next one. Cheers. Bye bye.