NEW Claude Agent View Just Dropped...

Claude's new agent view just drops. And if you run AI agents at all, this is the update you've been waiting for. Right now, most people are running Claude code and they're juggling five, six, even seven different terminal tabs at once. One might be doing keyword research.

One's writing content. One crashed 20 minutes ago and you don't even know it yet. You're spending more time managing the chaos than actually getting work done. Agent view solves that problem completely.

You get one window, every agent you're running, all visible at once. You can see what's active, what's finished, and what actually needs your attention without touching a single extra tab. And here's the part that blew my mind when I first tested this. You can now run fully autonomous agents in parallel or coordinate from one clean list whilst you do something else entirely.

I want to show you exactly how to set up, walk you through a real workflow I use every day, and show you one specific thing that you can do with agent view that makes your Claude agents way more powerful overnight. Stick with me to the end because the last thing I show you is the part that most people completely miss. Let's get into it. Today, we're going to be looking at the new Claude code agent view, which means you get one list of all your sessions and it's available as a research preview, right?

So you can run Claude agents to start dispatching multiple sessions at once, right? This is pretty cool because you don't need a terminal tab running to run this. You can just have an agent view and you can have these running together as separate sessions, but you can easily manage them with this, right? See, when you're running like agents in parallel before, you have to manage multiple terminal tabs or have a big grid of terminal agents, right?

It just got super messy, right? Whereas for example, if you're using agent view in Claude code, then you can manage everything at once and it's much easier to organize, right? So let's talk about how to use it. So you can press the left arrow from any session or you can run Claude agents from the terminal to open agent view.

Let's try this out. So this is inside the terminal, not inside the desktop app. Let's test this out now. And then we can start using this.

Let's try this out. I'm gonna create multiple sessions here and then we can start a task in the background, right? So once we've typed in Claude agent, we can then start running tasks in the background. So if we say, okay, keyword research for AI SEO, that's gonna start working here, right?

And so we can have multiple Claude code sessions running in the background rather than having separate terminals. So we can have separate terminals like this, but this gets quite messy, right? Sometimes you lose track. You don't really know what the name of these are, et cetera.

You don't really know what it's working on. Sometimes you forget to check it, et cetera. Now you can just have all of them running inside one single terminal window. So we've got this working on keyword research.

We can have another one creating a website. So let's say create a simple HTML landing page and now adds it to the working list, right? And so we've got multiple different Claude code agents working in parallel. We can see them with the agent view and we can see what's going on, right?

We can set up another one here. So create an AISEO calculator for organic keywords. And now it's working on those three and you can click on it or unclick on it. So if you want the full view of what's going on, you just click on them like that, right?

And then you press left to go back. So if we click on this one, these are three different tasks, as you can see. If we click on this keyword research task, we can go straight into that terminal for it and we can see what's going on, right? Now if we press left, then we can go back here and we can also see now that one task was completed.

So this website was completed all from the agent view. It's really crazy because you can deploy multiple tasks at the same time. You can be like way more productive, get way more done and it's all running in the background whilst I talk to you, right? And we can click on that and then we can go back to the task here.

And so the way that you manage agents is really changed forever because instead of having like multiple terminal tasks or for example, having multiple tabs or for example, setting up a new cloud session in each tab like that, which is super messy and difficult to manage, you can have them all running in one view, which is super useful. And then you can peek and reply without leaving. So you can select a session to peek at the last term. If it was waiting on a decision, you can click on it like that.

You can press left to go back and it's super easy. And if you want to see what's completed, you can see an example right here. It's really cool. Now this is something we use in auto mode as well to get stuff done.

So you can see we've got auto mode running, which just means like the agent can work autonomously and make decisions autonomously without me handling anything. And then you can see the task is ready to go here. So if we open this up, we've just created a HTML page just whilst I was talking to you. So that is how agents have really changed because previously you would have one by one or you'd be managing like five different terminal tabs across your screen.

Not a great way to manage your agents. Now you can manage them just in the one simple view like that. It's way easier. Now you might say, okay, why would you use this?

In terms of this, number one, you can scale the number of concurrent sessions. So you could dispatch several tasks and projects at the same time, each one paired with a skill and then return to a list of pull requests ready for review. That's an example, right? You could manage long running agents.

So you could actually have like agents managing your other agents in the background here. And you can easily navigate between separate sessions, which is just more productive, more easier, right? And you can see what's actually completed. So with the completed status here, we can see exactly what's happened and we can check them out and that sort of thing.

So really simple improvement, but it helps a lot and it just makes it much easier to manage your agents in one place. So thanks so much for watching. That's how to use the Claude Code Agent View, the new setup from Claude Code Agents. If you want to get more training on Claude and how to use it, et cetera, you go to the AI Profitable Boardroom, link in the comments description or go to aiprofitableboardroom.com, type in Claude and you can see we've got loads of different trainings on, for example, how to use Claude Code Agents, Claude Code Persistent Memory, how to use Claude Code Skills, et cetera.

And we have full tutorials and step-by-step guides on everything that you can see today. So if you really want to learn and master this stuff, we've got loads of training inside the AI Profitable Boardroom. Inside the community, you can ask questions, get help and support, et cetera. I create daily videos helping people as well.

Inside the classroom, you can get all of my new trainings and setup, including new daily advanced tutorials right here. So every day we just add in new advanced tutorials and showing you exactly what works, et cetera. We've video guides and step-by-step tutorials plus a hundred prompts on how to use each one. And you see all the trainings inside the classroom here.

So you get loads of cool stuff. Inside the calendar, you can jump on four weekly live coaching sessions. So you can jump on these live AI automation sessions, ask questions, get help and support, share your screen, meet other members who are doing similar things to you. And inside the map, you can connect with other people who are using AI automation and Claude to help you grow and scale, right?

It's just an awesome network of people who are doing amazing things. So feel free to check it out. Plus you can connect with me personally inside that too. Agents managing other agents like sessions connected.

Yes, absolutely, right? So if you go inside the terminal, you can have Claude agent view here, managing your sessions. Now we could create new tasks. It's okay.

Set up a new skill inside Hermes or AI SEO, right? And then that can just manage my other agents in the background. So we can go over here and it can start managing our Hermes agent, speaking to it directly using the Hermes agent MCP and then we're good to go, right? I'm actually gonna stop that because I don't want it to do that, but I just wanted to show you an example of how that works.

And then it'll be completed once done. And you just set up Claude agents to go back to it. What's the difference between Claude and Codex? If you look at Claude, it's more, it depends what you're comparing, right?

If you're comparing like the terminal usage like this, honestly, like Claude, I find is like a lot more useful for getting stuff done and being a lot more autonomous. Codex is good. It tends to run out of credits pretty quickly. I found even when I'm on the pay plan.

So that's one thing to be aware of. And also the Codex UI is totally different, right? It's totally different. So there's some big differences there.

They do similar things, but it just depends what's your favorite API. For me personally, I like to test and use them all because I like to learn about this stuff. What's the difference between Codex and Claude and Hermes agent? So Hermes agent is open source, whereas Codex and Claude are not open source, right?

Which means that you can customize Hermes agent as much as you want to. And also you're not stuck to one API with Claude and code and Hermes agent, right? With Hermes agent, you can switch between APIs. It's a lot more autonomous.

It can connect to your telegram and everything else. You can't really do that with Claude. You can connect it to your phone, but it's not very good, the phone connection, honestly. And also you just can't customize Claude in the same way, right?

And also Claude code is paid, right? Whereas for example, if you're using something like Hermes agent, you can use it for free because it's an open source free project. And then you can get free API to plug into it, like step 3.5 flash. Is there a way to allow Google Hermes to log into Google?

Yeah. So you can basically just give your login details to Hermes agent, and then it can use that email address. The other thing you can do is you can go to Google workspace, get an API, and then give that API with the authentication allowances to use Hermes agent. So Hermes agent can connect to email that way.

Is this the real Julian Goldie? It is indeed. Brian says, I would love a handover loop. So when context is masked, the agent does an automatic handover and the agents keep working with the new context.

So you can just, usually Claude will compact automatically if you need, if you run out of context, right? So you can just, for example, you can go into Claude's desktop like this and you just type forward slash compact, right? So if it ever runs out of context, which I've never really seen, I don't really see very often unless it's a huge task, but if you just inside Claude desktop or inside Claude code, you can run compact. And then that basically makes the context window smaller.

So it compacts into a small window. So that's the way that we're doing. How can I use Hermes agent to work on my mobile app? And how can I work with Hermes agent for marketing and so on?

So for me personally, I can use Hermes agent for editing videos, for creating SEO content, for doing daily blog content, et cetera, social media content. So those are some of the ways that you can use it for marketing. We have loads of training inside there I've got for boarding on that. And then when it comes to getting Hermes to work on your mobile app, you just go into Hermes setup and link them together, right?

So if you go into Hermes setup like this, we'll just enter on this section and then you go to the select platforms to configure section here, right? And that's how you can configure your channels that you want to use with Hermes agent. So if you want to link it to Telegram, just configure with Telegram right there. Oh yeah, I'm doing a huge task.

Compact can lose some details. So I usually get a handover document. Yeah, so what you can do is you can get, you can get all your agents to type any key notes as they go along into something like Obsidian, right? Obsidian is basically like a second brain.

So if you need to store context and notes, you can use Obsidian with your agents and it'll write down the notes inside a markdown file so that you never lose context again. So if you're finding like you're losing key details as your agents go along, for example, like Claude, what you can do is tell it as it works to know everything into Obsidian, write down the notes inside Obsidian, and then you've got all the key notes noted down as it works, right? And then also all your other agents can connect to Obsidian at the same time so that any context you have inside Claude gets automatically transferred to Hermes or Codex or whatever you want. Isn't Claude dropping third party configurations?

It depends what you mean by that. I don't know exactly what you mean, but I think they've dropped some stuff, but you can still connect easily to different things with the customize option, right? So if you go into Claude here and then you go to your settings like this, go to connectors and you can see they've moved it to customize and that's how you can connect to different MCPs, right? So for example, we've got Hermes MCP connected there, WordPress, GitHub, Claude and Chrome, et cetera.

That's how you can connect to third party applications. Okan says, so Hermes agent is better for coding and front end design than Codex and Claude? They're all pretty good at that, right? It just depends what you want to do.

You can use any of them, really. It depends what you prefer. If you prefer to manage AI agents from your phone, then Hermes agent is going to be better for coding and front end, right? If you're on your laptop all day and you prefer the UI of Codex and GPT 5.5's API, then you can switch to that.

So it really depends what you want to do. For me personally, I think the best for front end is Claude, but Hermes can do a pretty good job too.