ClawdBot Full Tutorial for Beginners: How to Use & Set up ClawdBot (Openclaw) transcript

Claudebot is changing the game when it comes to AI assistance. Most AI tools today live in a browser. You open a tab, you ask a question, you get an answer, and once you close that tab, everything just kind of disappears. They're useful, but they don't really feel like assistants. And that's exactly the problem Cloudbot is trying to solve. Unlike typical AI chat bots, Claudebot is a self-hosted AI assistant that runs directly on your computer.

It connects to powerful AI models like claude or chatgypt for intelligence while the agent itself lives locally on your PC. And this means it can actually do things for you like write code, create files, browse the web, and integrate with platforms like Telegram, Slack, and Discord. So, in this complete beginners tutorial, I'm going to walk you through everything that you would need to know about Claudebot. We'll cover what makes it special, whether you need specific hardware, the costs involved, and most importantly, I'll show you exactly how to set it up step by step from getting your API keys to having your first conversation through Telegram.

So, by the end of this video, you are going to have your own local AI assistant running on your own computer, ready to help you with your daily tasks. So, let's go ahead and dive in and get Cloudbot set up together. So, what exactly is Claudebot? Most AI tools today live in a browser. Again, you open a tab, ask a question, get an answer, and once you close that tab, everything just kind of disappears. They are useful, but they don't really feel like assistance at all.

They feel more like tools that you have to keep reopening over and over again. And that's exactly the problem Claudebot is trying to solve. Claudebot is a self-hosted AI assistant that runs directly on your own computer. So instead of living inside a web interface, it just connects to dedicated AI models like Claude or Chat GPT for its intelligence while the agent itself lives locally on your PC. What makes that interesting is what it can actually do.

And Claudebot isn't just limited to chat only. It can connect to Telegram, Slack, Discord, Google even, and of course other tools depending on how you set it up. You can use it to write code, build applications, do research, create lists, or handle tasks that you would normally bounce between multiple labs for. Because it's running locally, everything stays on your machine. Your conversations, your files, your context, none of that is sitting on some random third-party server.

And that's the real advantage here. So, with the right configuration, Cloudbot feels way less like a typical AI chatbot and a lot more like an actual assistant that understands your workflow and can interact with your computer. You can have it help build full projects, manage research, or automate things in the background all while chatting with it through Telegram or whatever platform that you prefer. So, what are the main features of Cloudbot?

When people hear AI assistant, they usually imagine something that understands their workflow and helps behind the scenes. [music] In reality, a lot of the tools stop at conversation and don't go much further than that. Once you start using Clawbot, though, it will be very much obvious that it's more than just a chatbot. It's designed to actually do things for you and not just respond to prompts. At a basic level, Claudebot can execute commands directly on your computer.

It can create files, write code, and even browse the web if you've configured it properly. And that opens the door to automating tasks and handling real work instead of just getting text responses. [music] It also integrates with messaging platforms like Telegram and Slack and Discord. So, you're not locked in to a single interface. You can talk to it from the tools you already use every day, which makes it feel a lot more natural to work with.

One of the more powerful features is memory. Claudebot can remember past conversations and your preferences over time. So, the more that you use it, the better it starts to understand how you work and what you usually need help with. On top of that, there's a web dashboard that gives you a full view of your chat history, the bot status, and all of your settings. So you don't need to dig through config files unless you actually want to.

Everything important is visible in one place. Cloudbot also supports multiple AI models. So if you have API keys for both cloud and chat GBT, you can just switch between them depending on what you need at the moment, which is great if you prefer one model for certain tasks over another. And finally, file creation is a big one here. Cloudbot can generate documents, spreadsheets, and presentations and other files too, directly from your requests without you having to open a single application.

Do you need a Mac Mini or a Mac device? A question that comes up a lot is whether you need a Mac Mini or even a Mac at all to run Cloudbot. And the short answer is nope. You can run Claudebot or Multibbot on pretty much any modern device, whether you're on a Mac, Windows, or Linux even. There's no special hardware requirements and no need to dedicate a machine just for this. It also doesn't use a ton of system resources because you're not running a full language model locally.

Claudebot just connects to existing AI models from providers like Anthropic or OpenAI. So your computer is mainly just running the agent itself. Are there costs involved? So when it comes to cost, Cloudbot itself is completely free all to use. You're not paying for the software or the agent. What you do need though is an API key from an AI provider like Anthropic or OpenAI. And that is where any cost comes from. How much you end up paying really does depend on how often you use it and which model you choose as your default.

More powerful or newer models of course will cost more per request, but they also tend to give better results. So there is a trade-off there. Lower tier models are cheaper, but the experience can feel a bit more limited. On top of that, some of Cloudbot's advanced features really do rely on additional thirdparty services, which means extra API keys and potentially extra cost. So, while Cloudbot itself is free, the total cost can vary quite a bit depending on how you configure it and how heavily you use it.

So, how difficult is it to set up Cloudbot? So, let's be real here. The setup isn't completely plugandplay. Sorry. Especially if you're not used to working with the terminal or dealing with API keys. There are a few moving parts and it can feel a bit intimidating at first. You'll need to install things like Node.js, add your API keys, and set up whichever messaging platform you want to use, like creating a Telegram bot.

Some of that involves running terminal commands and in a few cases editing JSON files, which can be unfamiliar territory if you're not a developer. The good news, though, is that once everything is configured properly, Cloudbot runs very smoothly, and you rarely need to touch the settings again. It's mostly just a one-time setup. The documentation definitely helps, but you may still run into a few bugs or some rough edges along the way, especially around API keys and gateway tokens.

But if you're comfortable following guides and using basic terminal commands, well, you should be able to get everything up and running in about 30 to 60 minutes. The hardest part here is usually just getting all the keys set up correctly and making sure your messaging platform connects properly. So, before we install anything, we do need to decide which AI model Claudebot is going to use. To do that, we'll need an API key from an AI provider.

I'm going to use Claude, but you can follow the same process with other supported providers if you do prefer. Claudebot supports multiple providers. So, if you'd rather use, say, OpenAI instead, you can just grab an API key from OpenAI and follow along with that. And if you're going with Claude, then head over to platform.claw.com and choose the option to get an API key. From here, you'll be asked to give the key a name.

And once it's created, all you got to do is just copy it. We'll be pasting this into Cloudbot during the setup in the next step. Now that we've got our API key, the next step is actually just installing Claudebot. And the process itself is pretty straightforward. On the CloudBot website here, you'll see a single line install command. Just copy that, open up your terminal, and paste it in. Once you run it, the installation will start automatically.

And at this point, you just let it do its thing. If Node.js JS or Git isn't already installed on your system, the installer will take care of that for you. In my case here, I already had everything set up since I am a bit of a developing developer, if you could say that. So, it's really just installing CloudBot itself. And once that's done, you just wait for the installation to finish before moving on to the configuration steps.

Once the installation finishes, Claudebot will automatically jump straight into the configuration process. Here, you don't have to manually start anything. It just picks up right where the install leaves off. Early on, you'll see a prompt that says something along the lines of, "I understand this is powerful and inherently risky. Continue." So, go ahead and say yes here so you can move forward with the setup. Next, you'll be asked to choose an onboarding mode.

So, for this setup in this video, go with quick start. It's the fastest and the easiest option and lets you get everything up and running without having to tweak a bunch of settings right away. After that, you're going to be asked to choose your AI provider. This is where you decide which model Claudebot will use for its intelligence. Since we grabbed a Claude API key earlier, I'm going to select Anthropic here. Just keep in mind this isn't a permanent decision.

You can use any supported provider that you want and it really comes down to personal preference. After selecting Enthropic, choose Enthropic API key and paste in the key you copied earlier. After that's set, Cladbot will ask you to pick a default model. There are quite a few models available, so this part is really about balance. In my case, I went with Sonnet 4.0 because it's efficient and still has a fairly recent knowledge base.

More powerful or newer models will generally give better results, but they will also burn through your credits a lot faster, and that's something to keep in mind here. The good thing is you're not locked into this choice. You can always switch models later if you just want to. Just be aware that each model has its own personality, so to speak. So, if you change it down the line, the responses might be a little different.

So, after that initial setup in the terminal, the next step is moving the conversation over to Telegram. Open up the Telegram bot that we just created and send the forward/start command to kick things off. After a short moment, Claudebot should respond, letting you know that the connection is working and that it's ready to chat with you. Once you see that message, you're officially talking to your own local AI through Telegram.

And from here, you can start having real conversations with it. I'm going to ask it a few questions, bring up some general topics, and then give it more information about myself and what I do. And you can see that it responds properly and keeps the context of the conversation. This part is actually pretty important because the more you talk to Claudebot and give it context about how you work, the better it gets at understanding you.

And just like with training AI models, giving it more information over time helps it build better memory and respond in a way that fits your workflow. Also, let's go into the Claudebot dashboard here where you can see the chat history showing up in real time. Now, that's a good way to confirm that everything is connected correctly and that's all working as expected. After setting everything up and chatting with Cloudbot just for a little bit, you might start wondering, huh, what else can it actually do?

Well, Telegram is a great starting point, but it's really just the beginning here. Cloudbot isn't limited to Telegram only. It can also integrate with other platforms like Slack, Discord, and additional messaging tools. You just need to configure them in the settings. Once set up, you can interact with the same local AI from multiple places. It can also interact directly with your computer. Cloudbot can run bash commands, create files, and help automate tasks, which makes it especially useful for things like coding, file management, or repetitive workflows.

So, if you connect a web search provider like say Brave Search, Cloudbot can pull in up-to-date information from the web instead of relying only on its training data. That's really useful when you need current information or just some quick research. Memory is another big part of how Cloudbot works. It remembers past conversations and can reference them later. So, the more you do actually talk to it, the better it understands your preferences and how you work.

On top of all of that, it can also write and run code directly. Why not? Of course, [music] generate documents, spreadsheets, presentations, and manage files, all from chat. You can ask it to create reports, organize your data, or draft documents without opening any other apps. You can also customize how Cloudbot behaves, what it calls itself, and how it responds. So, it does feel more tailored to your workflow instead of a generic assistant.

Also, there's a built-in web dashboard at localhost col18789 where you can see your full chat history, monitor the bot status, and manage settings without ever touching configuration files. And finally, everything is privacy focused. Because CloudBot runs locally on your machine only and connects directly to the AI provers API, there's no routing your conversations through third-party servers. At the end of the day, Claudebot is basically your own local AI powerhouse.

Something that you can talk to from your phone or your desktop, but that actually lives on your machine and works the way you want it to. And that's pretty much it. That's CloudBot in a nutshell. It's a local AI assistant that actually works with you, not just talks to you. We focused on the core setup here so you can get up and running without over complicating things. But there is a lot more you can build on top of this once you are comfortable.

Thanks for watching and investing your time with me today. I'll see you at the next one.