OpenClaw Multi-Agent Setup in Slack (Pazi AI)

7:59

OpenClaw Multi-Agent Setup in Slack (Pazi AI)

Mervin Praison 08.05.2026 753 просмотров 15 лайков

Machine-readable: Markdown · JSON API · Site index

Смотреть на YouTube

Поделиться Telegram VK Бот

Транскрипт Скачать .md

Анализ с AI

Описание видео

A 24/7 agentic team in your Slack, deploying real apps to Google Cloud. Manager agent delegates, developer agent builds, DevOps agent ships — all from one channel. Pazzy.ai: https://pazi.ai/ In this video I show you how to set up multiple OpenClaw agents (manager, developer, DevOps), connect each one to Slack in three steps using Pazzy.ai, and run a real end-to-end task: building a Flask app and deploying it to Google Cloud Run — with the agents tagging each other and reporting status as they work. Thanks to Pazzy.ai for sponsoring this video. Try Pazzy.ai: [your affiliate link] What you'll learn: - Creating multiple agents in OpenClaw via the CLI - Configuring agents through Pazzy.ai's interface (no terminal headaches) - Connecting each agent to Slack with manifest + app token + bot token - Setting up manager → developer → DevOps coordination with shared memory - Authenticating Google Cloud CLI inside an agent - Asking the team to build and deploy a Flask app from a single Slack message - Tagging specific agents mid-task to refine work in real time Trade-offs to know: - Configuring agents directly in OpenClaw works but involves several terminal commands - Slack integration through Pazzy.ai was the smoothest of the options I tested - Service account keys are the recommended auth path; CLI login is faster for demos Chapters: 0:00 Multi-agent team in Slack — what you're getting 0:58 Adding agents in OpenClaw via CLI 1:21 Why I'm using Pazzy.ai for setup 2:20 Creating the manager agent 2:35 Connecting the agent to Slack (3 steps) 3:37 Adding the DevOps and developer agents 4:08 Adding the bots to a Slack channel 4:28 Giving the manager agent its instructions 5:07 Authenticating Google Cloud for the DevOps agent 6:09 Giving the team the real task 6:34 Manager → Developer → DevOps in action 7:24 Final deployment and the live URL - OpenClaw: https://docs.openclaw.ai/concepts/multi-agent Subscribe for more agent workflows, AI tools, and honest builds. #OpenClaw #AIAgents #Slack This video showcases the power of OpenClaw's multi agent systems, demonstrating how a manager agent orchestrates tasks. Watch as AI agents communicate and report through Slack, enabling seamless ai automation and deployment to GCP. This setup highlights the practical use of ai tools for continuous operation and efficient task management.

Оглавление (12 сегментов)

Multi-agent team in Slack — what you're getting

Openclaw with multi-agent setup and linking all these agents directly in Slack. And now I have a full-fledged agentic team and I can ask it to perform any task at any time 24/7. I asked the manager agent create a basic test app and deploy in Google Cloud. And the manager agent is tagging the developer agent and giving the task information. The manager agent asking to improve few more features. Once that is completed, the manager agent asks the DevOps agent to deploy the application. The DevOps agent deploys that to Google Cloud and gives the production URL. That is a published URL. And I'm going to take you through step-by-step how you can create multiple agents, integrating that with Slack, and deploying it in cloud servers such as Google Cloud or any other cloud. That's exactly what we're going to see today. Let's get started. —

Adding agents in OpenClaw via CLI

— Openclaw has the ability to create multiple agents. By default, it comes with one agent. But to add multiple agents, this is the command. In your terminal, just type openclaw agents and add and the agent name, that is manager, and that will automatically create the manager agent for you. The agent manager ready. You can see the agent manager here. But configuring it directly in

Why I'm using Pazzy.ai for setup

Openclaw is a bit complicated and you need to run all these terminal commands on your own. That makes it complicated. To solve that, we have Paz. ai. Using this tool, you can configure all your agents in this easy-to-use interface. And for each agent, you can connect it to Slack directly from here. And I found Slack integration through this is more easy compared to any other tools I've tested. And behind this, Openclaw is running. And you don't need to manage your own Openclaw server. Everything is managed here. And you can get started for free. You can build your 24/7 super intelligence team with DevOps, developer, QA, and many other members right from your Slack or any other tools such as Telegram, WhatsApp, and much more. Thanks to Pass AI for sponsoring this video. Initial task given to the manager agent, then it manages the developer agent and the DevOps agent to

Creating the manager agent

complete the task. So, once after you sign up, you can create your new agent by clicking this new agent button. There you got a interface to give you a name for your agent. I'm going to name it as manager agent, and then create agent.

Connecting the agent to Slack (3 steps)

Here is the manager agent. I can connect to Slack directly from here, and the option is very easy. Only three steps. First, go to this URL, create new app from manifest. There you choose Presenter AI, and then click next. Now, here is where you paste your manifest content. So, here is your manifest content. Just copy that, and then paste it here, and then click next, and then create agent. Now, here is the manager agent, and we need to generate the token. So, generate token and scopes. I'm naming it as manager agent, and then adding a scope. It's connections right. That's it. Now, generate. Now, you got the app token, which I'm going to copy, and I can add that app token here. Next, I need bot token. Again, it's easy. Coming back to the page, going at the top, install app, and then install to the workspace, and authorizing it. Now, you got the bot token ready, and it's integrated with Slack. I'm going to add the bot token here, and then connect

Adding the DevOps and developer agents

to Slack. That's it. Now, your Slack is integrated. Similarly, even if you have any issues, you can ask the manager agent if you find any issues along the way, or even if you want to start it from scratch. Similarly, I'm going to create other agents, got the DevOps agent, then the developer agent, all ready and connected to Slack. And I feel that this is easier way to integrate to your Slack, and it'll automatically restart your gateway. Coming back to Slack, you can add your apps when you go under the apps column. There you'll find

Adding the bots to a Slack channel

all the agents such as manager agent, developer agent, and much more. Now, I'm going to open one of my channel. There I can add my bots, such as the DevOps bot, developer bot, and the manager agent. So, now you can see automatically all these agents are responding. And we got the response here. Now, to

Giving the manager agent its instructions

the manager agent we can tell, "Your work is to get the task and then pass it on to relevant DevOps agent and developer agent. " Now, it's reviewing it, and I'm going to ask you to save it to its memory. "Save it in your memory. " Also, while responding, tag agent @DevOpsAgent and developer agent. "DevOps agent, on completing the task, respond back to the manager agent regarding the status. " Similarly, developer agent. I'm going to click enter. Now, you can see all the agents understanding their respective

Authenticating Google Cloud for the DevOps agent

tasks, and I've got the response here. And everything is ready to receive tasks. But one final thing to do, I need to provide Google Cloud access to the DevOps agent. So, how I'm going to do that? I can even ask the DevOps agent from Slack itself. Another option, go to Pazy and then choose the DevOps agent. There I can chat with the agent, install G Cloud, that is Google Cloud CLI, and then configure the authentication. It gave me three options. One is G Cloud interactive login. Second option, service account key, that is a recommended option. And third, workload identity. Go to Google Cloud console and create a service account and assign all the required permission and provide the key data to this agent. But to make it much more quicker, I'm going to ask you to log in like this. You do option one and give me the URL, I will authenticate. And it's giving me the authentication URL and I'm authenticating. And that will provide me the authentication code. Going to copy this and paste it here, then I'm in. Now

Giving the team the real task

it's all authenticated. You can also chat with the DevOps agent directly from here, too. Just to show you that you got two different to talk to the same agent. Now I'm going to give a task to the manager agent. Calling manager agent. Create a dedicated Flask application with UI and deploy in Cloud Run. That is in Google Cloud. And then clicking enter. Now you can see the manager agent is

Manager → Developer → DevOps in action

calling the developer agent and asking you to create a Flask application with front end. And now you can see the developer agent is working. Now it completed a task and informing the manager agent. Now the manager agent is responding to the DevOps agent to deploy the application. And the manager agent is updating me regarding the status. The developer agent has finished the Flask app. I've now handed off to DevOps. Now you can see the DevOps agent is working. Even while these agents are working, I can even tag any agent to perform a task. Even if you want to improve the user interface, I can tag a specific agent and make it perform a task. Or I can even inform the manager agent to optimize the workflow. I really like this collaboration directly within your Slack. And finally

Final deployment and the live URL

the DevOps agent completed a task and I've got the final URL, which I'm going to open. And the manager agent is giving me the report. And here is the application which got created. A simple application, but you can see all the agents work together having Open Cloud as the back end, and Paz AI making it simple. And now I have a full-fledged agentic team, and I can ask it to perform any task at any time, 24/7. Do let me know in the comments below what you think about this. I hope you like this video. Do like, share, and subscribe, and thanks for watching.

Другие видео автора — Mervin Praison

Ctrl+V

Экстракт Знаний в Telegram

Экстракты и дистилляты из лучших YouTube-каналов — сразу после публикации.

Подписаться

Лучшие методички за неделю — каждый понедельник