OpenClaw + Kilo CLI: Deploy AI Coding Agent That Runs 24/7 (Clawdbot)

Hey everyone, welcome back with me Nathan. In this video, I want to share

Nathan. In this video, I want to share how you can use Open Cloud and Kilo CLI to deploy autonomous AI coding agent that assists you in building software projects. If you need a quick refresher,

projects. If you need a quick refresher, OpenCloud is basically a self-hosted agent runtime that lets you connect chat apps like WhatsApp, Telegram, and more

to AI coding agents. It acts as a personal AI assistant that can do things on your machine equipped with tool use, sessions, memory, and multi- aent routing. For example, over here,

routing. For example, over here, OpenCloud helps the user build an entire app just by typing a prompt through WhatsApp. It creates a functional SAS,

WhatsApp. It creates a functional SAS, then keeps refining it, fixes errors, and iterates until the project is completed. Open Cloud can be connected

completed. Open Cloud can be connected with all kinds of tools. And in this video, I want to share how you can integrate Kilo CLI with Open Cloud. Kilo

CLI is an open-source agendic terminal tool that can help you plan and build your software projects. It also enables you to orchestrate multiple sub agents all directly from the terminal. It

supports over 500 different AI models so you can pick the best for your use case.

Sometimes Kilo also teams up with AI providers to let you use models for free. For example, at the time of this

free. For example, at the time of this recording, you can use Miniax M2.1, GLM 4.7, and a few other stat models without incurring any charges. By combining open

cloud and kilo CLI, you can have a fully autonomous coding agent that you can reach from chat apps anytime. You can

ask open cloud to plan, build, debug, and review your projects by just chatting and it will plan and execute tasks using Kilo CLI. Next, let me walk you through exactly how to set this up

from scratch. Now, before we jump into

from scratch. Now, before we jump into the exciting part, don't forget to subscribe to my YouTube channel so you won't miss any new videos that can really level up your skills. Also, click

on the bell to get notified whenever a new video comes out. Thanks a lot for supporting the channel. You really help me make useful videos just like this one. First, if you don't already have

one. First, if you don't already have Node.js installed, then head over to node.js.org and then install the version

node.js.org and then install the version for your operating system. Once you have NodeJS installed, you need to install the Kilo CLI. So, just run the npm install command as instructed in the

documentation. It will take a few

documentation. It will take a few seconds for the installation to complete. And when it's completed, you

complete. And when it's completed, you can verify the installation by running the kilo command and it should launch the interactive terminal session. Next,

run the /connect command to connect your AI provider to the CLI. Here, I'm using the Kilo gateway, so I just need to authorize the CLI. Once connected to a provider, you can then select the AI

model you want to use with the tool.

There is the Minimax M2 model, which is currently available for free. There is

also the GLM 4.7 model that offers a free tier. So just pick any model you

free tier. So just pick any model you want here. Once you select the AI model,

want here. Once you select the AI model, you can test the connection by sending a prompt. I will send a simple hello prom

prompt. I will send a simple hello prom here. And we can see the reply from the

here. And we can see the reply from the AI model. So everything is working as

AI model. So everything is working as expected. Next, let's install OpenCloud.

expected. Next, let's install OpenCloud.

On the documentation page, you can see the bash script that you can run from the terminal. But OpenCloud also has

the terminal. But OpenCloud also has this npm install command. So let's just use that instead. I will open a new terminal window. Paste the command and

terminal window. Paste the command and then press enter. It will take some time for OpenCloud to install. So I will fast forward to when it's finished. Okay. Now

OpenCloud has been installed on the computer and the next step is to run the OpenCloud on board command which will guide you on the setup. First you need to acknowledge that running OpenCloud is

powerful and inherently risky. If you

scroll up a bit, there is also a security warning here. It also links to the security documentation page. So it's

recommended to check it out first. Next,

select yes to continue and then choose quick start for the easiest setup. Now

you need to choose which AI model that will power your open cloud. You have

several options here. The most popular AI models are all available. For

example, select entropic here and then you can choose whether to authenticate with all token or API key. Next, you can select messaging channels that you want to connect with open cloud. Let's skip

this step for now as I will show you how to add a messaging app later. After

that, OpenCloud will ask if you want to enable any skills. This is optional, so let's just do that later. Finally, it

will install the gateway service that will serve as the main connection point between OpenCloud, your computer, and all connected messaging channels. Once

done, you will be asked where you want to hatch the bot, whether in the terminal or in web dashboard. Let's

select web dashboard over here as it has better UI. This will open the web

better UI. This will open the web dashboard directing you to the chat interface. You can see here that the

interface. You can see here that the assistant is finally online. These are

all automated messages. First, the

OpenCloud bot checks on existing files.

There are several core markdown files used by the bot to keep track of itself, you as the user, the automation, and tasks it will be designed for. You can

learn more of these concepts in the documentation, but for now, the AI is looking at a setup file and realize it's a blank slate. So it asks you to configure the bot giving it a nickname

telling it about who the user is and so on. I will simply say you are Jarvis.

on. I will simply say you are Jarvis.

I'm Nathan and we're going to work on software projects together. Send the

prom in. The AI will then update the core files and after a while the AI is ready to work asking for the first project. All right. Now that the open

project. All right. Now that the open cloud bot is online the next step is to add a skill that will teach it how to use Kilo CLI. You can look for the skill in cloud hub which is the public skills

registry for open cloud. Over here I have created the kilo cli skill that provides instructions on how to run kilo cli to perform code changes, reviews,

fixes and create pull requests. To

install it just copy the URL here and then back in open cloud as it to install the skill and provide the link to it.

After a few seconds the bot will install the skill using cloud hub. You can see the path to the skill here. And that's

it. With the skill installed, OpenCloud now knows how to run Kilo CLI. Next, we

need a project to work on. So over here, I have prepared a web app project. This

is a customer relationship management dashboard created using Nex.js, TypeScript, and Tailwind CSS. If we open the agents.md file, we can see the full

the agents.md file, we can see the full text stack. It uses Nex.js, TypeScript,

text stack. It uses Nex.js, TypeScript, Telwin, Drizzle for OM, and SQLite for database. We can take a look on the

database. We can take a look on the browser. Here's the dashboard. There are

browser. Here's the dashboard. There are

charts and insights and the customers page here is already created, but other pages like the deals page over here is not yet completed. So, we're going to ask OpenCloud to work on this project.

First, run the pwd command to print the path to the project. We need to copy this path and then paste it into OpenCloud so that it knows where to find the project. I will add the prompt here

the project. I will add the prompt here asking OpenCloud to use kilo CLI on the project. Build the deals page. Let users

project. Build the deals page. Let users

create, view, update, and delete the deals. Reuse as much existing components

deals. Reuse as much existing components as possible. Okay. Send the prompt in

as possible. Okay. Send the prompt in and let the bot work on the request. It

will take a few minutes to fulfill this request. So, I will skip a bit to when

request. So, I will skip a bit to when it's finished. All right. Over here,

it's finished. All right. Over here,

OpenCloud has finished the request. We

can click on this shell execution. And

we can see the message OpenCloud used T-Max to run a kilo instance. And if we scroll up, we can see all the progress being done. And then here, OpenCloud

being done. And then here, OpenCloud executed the kilo run command as instructed in the skill. If we open the project in VS Code, we can see the changes being made to the deals

component. And then if we open the

component. And then if we open the browser, we can see the newly created deals page. It looks nice and I will

deals page. It looks nice and I will review this later. But for now, you can see how you can instruct OpenCloud to use Kilo CLI to help you develop your app projects. Now to complete the

app projects. Now to complete the workflow, we need to connect a messaging channel to OpenCloud so that you can reach it from outside the dashboard. In

the documentation page, there are instructions on connecting popular messaging apps like Slack, Discord, WhatsApp, and many other. But for this demo, I will show you how to connect

Telegram as it's the easiest to set up.

First, open a new terminal window and then run the Open Cloud Configure command. This command allows you to

command. This command allows you to change OpenCloud configuration. Select

the local gateway and then it will ask you what to configure. So select

channels and then the configure or link option. On the next option, select

option. On the next option, select telegram. And now we need to create a

telegram. And now we need to create a new telegram bot and paste the API token over here. So head over to telegram. I

over here. So head over to telegram. I

have it open on my iPhone here. And then

chat with the bot father. You can search for the bot account. Make sure that you select this one with the verified batch.

Next, tap the create a new bot button here. Now enter your bot detail. I will

here. Now enter your bot detail. I will

just name this as Nathan opencloud bot.

Also the same for the telegram URL here.

Tap create bot and here we can see the API token to connect this bot. So just

copy that and then paste it into open cloud here and that's it. Now select the finished option and now it asks us if you want to configure the DM access policies. Uh the default pairing is

policies. Uh the default pairing is fine. So let's select no. Now we can

fine. So let's select no. Now we can continue and then the configure session will be completed. Go back to open cloud dashboard and we can see the telegram bot is now connected to open cloud but

we still need to allow the user to send message to the telegram bot. So back in telegram tap on the bot you created and then start a chat with it. It will say

open cloud access is not configured and then show your telegram user ID and pairing code. So back in open cloud

pairing code. So back in open cloud scroll down until you see the allow from section as shown here. Click the add button and then copy the telegram user ID. I will open it on the browser here

ID. I will open it on the browser here so that I can copy it and just paste the telegram user ID in the allow from input box. Now scroll down until you see the

box. Now scroll down until you see the button to save telegram configuration.

Click that and now you should be able to chat with your telegram bot. So here I will send a message to the telegram bot.

Just say hey.

And yep we can see the response from the bot over here. It says the deals page is done. and working and what's next. So

done. and working and what's next. So

this means the telegram bot is now connected to open cloud and you can simply send messages from telegram whenever you want something to be done on your projects. So yeah that is

basically the full setup for using open cloud with kilo cli. To recap, we installed kilo cli globally using npm and then set up open cloud with the onboarding wizard. Added the kilo cli

onboarding wizard. Added the kilo cli coding agent skill from cloud hub so that opencloud knows how to delegate coding tasks to kilo. This gives you a fully autonomous coding agent that you

can control from anywhere from your desktop, your phone, or even through automated schedules. And since Kilo

automated schedules. And since Kilo supports over 500 models, you have the flexibility to use whatever model workspace for your use case. If you want to learn more about Kilo CLI, you can

visit the documentation as it shows everything you can do with the tool. And

that brings us to the end of this video.

Now you know how to use Open Claw and Kilo CLI to deploy your coding agents that can help you build software on the go. If you're new to the channel, my

go. If you're new to the channel, my name is Nathan and I help you learn how to code and use AI tools. Make sure to subscribe if that's something you find useful. Don't forget to turn on the

useful. Don't forget to turn on the notification bell, like this video, all the good stuff as it really helps this channel to grow. With that being said, thanks so much for watching until the end and I'll see you in other videos.

Bye.

[music] >> [music]