Hermes Agent Kanban Swarms are INSANE!

**Julian Goldie** (0:00)
Today, we have a brand new update from Hermes agent on Kanban boards. So this is a brand new update that basically means that you can build more and better multi agent Kanban teams. And I'll show you some examples of what we've done with this in a second. And you can have larger concurrent sessions and more with this new update. Now you can see the tweet right here. I want to break down exactly what this means step by step so that if you're not sure how to use Kanban boards or if you want to indicate what are the actual use cases for having teams of agents orchestrated with a Kanban board. Let me show you exactly how this works. First thing you want to do is just make sure that you actually update your Hermes agent, which you can do inside your terminal. Just go to the Hermes update or you can go to the manage section inside your Hermes dashboard. And from here, scroll down and click update Hermes right here.
Now, once you've done that, you'll get access to the newer version of Hermes. And let me show you exactly how this works and what it means. So line by line, scaling huge multi-agent Kanban runs. This is a headline. So the one thing that used to break, which was running a whole board of agents at once, now actually works.
And I'll show you exactly how we did this with something called the content factory. Also, you might be wanting to create large concurrent sessions. What does that mean? Basically, you can have loads of agents or chats actually running at the same time without breaking Hermes agent. So you can have more agents running together in parallel, which means more agents and less freezing, which is great. Also, there's one fix for this. So this new update to database concurrency, they say, basically slow saves stopped freezing everyone. I explained exactly what that means in a second. And then just make sure that you updated to the latest version to get access to this new version of Hermes and how to use it. And the main point here is like you don't need a smarter AI, you just need agents to don't trip over each other. And that's all this update fixes, but it's pretty huge. So what does this mean in plain English? Well, basically imagine you hire like 10 AI workers to run everything for you. They might share one assistant and that assistant writes down everything inside a single notebook. Here's the problem you keep hitting. Every time one worker needs something saved, the other nine have to wait their turn at the notebook. And if one save is slow, everyone freezes until it's finished. That was the old way. So the second you try to run more than one or two agents, the whole thing would just break. This update basically changes that so that the assistant, for example, would now hand slow saves to a back office helper, so your workers never stand around waiting on each other. And the result is you can finally run a whole team of agents once. So for example, that could be like a research team, whole agency of agents, could be a content factory and it doesn't freeze. That's the update. So let me talk through exactly what we've built with this and how it works. And this is something I call the unfreezeable loop. So every agent system that scales really follows the same five news.
So basically you find out, okay, what's freezing your agents, you move slow jobs out of the way, you keep the team moving, you stop two agents doing the same job, and you lock it in so it can't break again.
Now, if we look at this one, for example, you can actually run like a team's pulse.
And you can see here that basically it would just run on a loop with all your agents getting stuff done. Now, if you want to indicate how does that work in reality, we actually built something called a content factory inside our agent OS. And here's a real use case for this. So if we go to the Kanban board section over here, this is our agent Kanban. Now we've got multiple boards with multiple teams inside each board. And you can see how many different agent profiles we have. Now what we can actually do is we can dispatch our agents to work on a task, and it will be assigned to relevant teams of agents across each Kanban board. And so each board, you could have like a separate project. So it could be, for example, like a video agent on one board. It could be, for example, creating a game on another board. And then you could have, for example, like a website you create on another board. And then you could have multiple agent profiles, like for example, we have a SEO writer, SEO backlink builder, SEO keyword generator, SEO outline writer, all working together inside one task. Now, as soon as we drop a task into the Kanban, it will be automatically triaged by a head agent. And then that will be sent to the different sections, depending on where the task is at. So if it's in triage, obviously it will stay here. If it's actually ready to go, it's going to be in the ready section. If it's completely finished, it will show up in the done section. That's the difference here.