I Tried Herdr Instead of tmux, Here’s What I Found
Read full transcript 10 segments
-
Hey, what's up, guys? I've been trying Hey, what's up, guys? I've been trying out Herder for the past few weeks and out Herder for the past few weeks and out Herder for the past few weeks and I'm really liking it so far. So, in this I'm really liking it so far. So, in this I'm really liking it so far. So, in this video, I'm just going to show you how I video, I'm just going to show you how I video, I'm just going to show you how I use it and who should maybe use it and use it and who should maybe use it and use it and who should maybe use it and maybe who shouldn't use it. So, Herder, maybe who shouldn't use it. So, Herder, maybe who shouldn't use it. So, Herder, this came out a few months ago and at this came out a few months ago and at this came out a few months ago and at the time I was thinking, "Oh, that's an the time I was thinking, "Oh, that's an the time I was thinking, "Oh, that's an interesting concept." Since then, it did interesting concept." Since then, it did interesting concept." Since then, it did now has 40,000 stars on GitHub, so now has 40,000 stars on GitHub, so now has 40,000 stars on GitHub, so that's always a good sign. There is very that's always a good sign. There is very that's always a good sign. There is very active development going on and they active development going on and they active development going on and they also have recently gotten some funding. also have recently gotten some funding. also have recently gotten some funding. So, So, So, uh let's see. Do I have uh let's see. Do I have uh let's see. Do I have um um um No, I don't think I can see that here, No, I don't think I can see that here, No, I don't think I can see that here, but in any case, I I read somewhere that but in any case, I I read somewhere that but in any case, I I read somewhere that they raised some money in a seeding they raised some money in a seeding they raised some money in a seeding round. So, they're probably going to round. So, they're probably going to round. So, they're probably going to break create some sort of SaaS service break create some sort of SaaS service break create some sort of SaaS service out of this with a cloud solution, which out of this with a cloud solution, which out of this with a cloud solution, which is great. Like, if if you need that, is great. Like, if if you need that, is great. Like, if if you need that, that's great. So, Herder is basically that's great. So, Herder is basically that's great. So, Herder is basically uh tmux for AI agents. And the reason uh tmux for AI agents. And the reason uh tmux for AI agents. And the reason why I was a little bit skeptical at why I was a little bit skeptical at why I was a little bit skeptical at first was because first was because first was because what it does is it it keeps track of the what it does is it it keeps track of the what it does is it it keeps track of the status of your AI agents and then you're status of your AI agents and then you're status of your AI agents and then you're able to um switch between them, etc.
-
able to um switch between them, etc. able to um switch between them, etc. And I was already doing this inside of And I was already doing this inside of And I was already doing this inside of tmux. So, if you have been watching any tmux. So, if you have been watching any tmux. So, if you have been watching any of my content before, then of my content before, then of my content before, then you know that I'm a big fan of tmux and you know that I'm a big fan of tmux and you know that I'm a big fan of tmux and I always have these tabs in the top here I always have these tabs in the top here I always have these tabs in the top here and I had built some integrations that and I had built some integrations that and I had built some integrations that would change my title here. So, right would change my title here. So, right would change my title here. So, right now it's set, so it changes the title uh now it's set, so it changes the title uh now it's set, so it changes the title uh based on the current directory in the based on the current directory in the based on the current directory in the top left here. top left here. top left here. But, I I figured out how I could use But, I I figured out how I could use But, I I figured out how I could use that. So, when my Claude agent was that. So, when my Claude agent was that. So, when my Claude agent was finished working, then it would have finished working, then it would have finished working, then it would have like a checkmark there. So, that worked like a checkmark there. So, that worked like a checkmark there. So, that worked pretty that worked well, that was all I pretty that worked well, that was all I pretty that worked well, that was all I needed. But, then um I made some needed. But, then um I made some needed. But, then um I made some updates, I changed the diff to a updates, I changed the diff to a updates, I changed the diff to a different system, and then I had to different system, and then I had to different system, and then I had to basically configure that again. I was basically configure that again. I was basically configure that again. I was using a tmux socket for that. using a tmux socket for that. using a tmux socket for that. And I was I wanted to do this more And I was I wanted to do this more And I was I wanted to do this more reliably over SSH. reliably over SSH. reliably over SSH. And then I figured, well, let's just And then I figured, well, let's just And then I figured, well, let's just let's just try out Herder. And so far I let's just try out Herder. And so far I let's just try out Herder. And so far I actually really like it. The way I work actually really like it. The way I work actually really like it. The way I work with it now is that I basically have my with it now is that I basically have my with it now is that I basically have my AI agent terminal window, like so so AI agent terminal window, like so so AI agent terminal window, like so so that's this one.
-
that's this one. that's this one. This is where I have Herder open. And This is where I have Herder open. And This is where I have Herder open. And then for my normal command line work for then for my normal command line work for then for my normal command line work for all of my other all of the other stuff all of my other all of the other stuff all of my other all of the other stuff that I'm doing, I just have a second that I'm doing, I just have a second that I'm doing, I just have a second terminal open with my usual tmux, etc. terminal open with my usual tmux, etc. terminal open with my usual tmux, etc. So I have a dedicated workspace command So I have a dedicated workspace command So I have a dedicated workspace command line workspace for working on my AI line workspace for working on my AI line workspace for working on my AI agents. So this is not running on my agents. So this is not running on my agents. So this is not running on my local machine. If I do Herder machine local machine. If I do Herder machine local machine. If I do Herder machine list, list, list, then you will see that I have one then you will see that I have one then you will see that I have one machine Mercurius, that is my laptop machine Mercurius, that is my laptop machine Mercurius, that is my laptop actually, and I'm recording this on my actually, and I'm recording this on my actually, and I'm recording this on my Framework desktop. But if I run Herder Framework desktop. But if I run Herder Framework desktop. But if I run Herder now, then it's going to connect what's now, then it's going to connect what's now, then it's going to connect what's going to start up, then you have my going to start up, then you have my going to start up, then you have my local instance here, but it's also local instance here, but it's also local instance here, but it's also connecting to my other machine. And I connecting to my other machine. And I connecting to my other machine. And I really like this idea because currently really like this idea because currently really like this idea because currently I have this running on my own laptop, so I have this running on my own laptop, so I have this running on my own laptop, so I can carry it with me, but I can also I can carry it with me, but I can also I can carry it with me, but I can also have one or more cloud servers that are have one or more cloud servers that are have one or more cloud servers that are added in this list that all can be added in this list that all can be added in this list that all can be running AI agents for me. So that's a running AI agents for me. So that's a running AI agents for me. So that's a one really neat feature that you one really neat feature that you one really neat feature that you basically have this one single pane of basically have this one single pane of basically have this one single pane of glass where you can manage all of your glass where you can manage all of your glass where you can manage all of your AI agent terminals, even across multiple AI agent terminals, even across multiple AI agent terminals, even across multiple machines. That is something that I machines. That is something that I machines. That is something that I really like.
-
really like. really like. The second thing then is that The second thing then is that The second thing then is that it it it basically takes on a lot of tmux basically takes on a lot of tmux basically takes on a lot of tmux conventions. So if I do control B C, I conventions. So if I do control B C, I conventions. So if I do control B C, I can create a new tab like this, and I I can create a new tab like this, and I I can create a new tab like this, and I I basically now just have basically now just have basically now just have a terminal on that other machine. So a terminal on that other machine. So a terminal on that other machine. So that like I could SSH over there, but if that like I could SSH over there, but if that like I could SSH over there, but if I just need to run a quick command on I just need to run a quick command on I just need to run a quick command on that machine, it's really useful for me that machine, it's really useful for me that machine, it's really useful for me that I can just open it like this and that I can just open it like this and that I can just open it like this and just run a quick command and and do my just run a quick command and and do my just run a quick command and and do my thing there. So, I am currently running thing there. So, I am currently running thing there. So, I am currently running three to four AI agents at the same three to four AI agents at the same three to four AI agents at the same time. time. time. So, you can click around it. You can So, you can click around it. You can So, you can click around it. You can click around with your mouse. So, these click around with your mouse. So, these click around with your mouse. So, these are all working at the moment. They are are all working at the moment. They are are all working at the moment. They are doing their things. But, doing their things. But, doing their things. But, it is now waiting on a review agent, for it is now waiting on a review agent, for it is now waiting on a review agent, for example, and when that will be when that example, and when that will be when that example, and when that will be when that is done, it is going to show that it's is done, it is going to show that it's is done, it is going to show that it's ready here when this bubble over here on ready here when this bubble over here on ready here when this bubble over here on the left side will go green. And also, the left side will go green. And also, the left side will go green. And also, it plays a little sound. And that is it plays a little sound. And that is it plays a little sound. And that is that was something I didn't know I that was something I didn't know I that was something I didn't know I missed until it started happening.
-
missed until it started happening. missed until it started happening. In my own tmux setup, I mean, I could In my own tmux setup, I mean, I could In my own tmux setup, I mean, I could have created that, but have created that, but have created that, but I didn't have that, but it's actually I didn't have that, but it's actually I didn't have that, but it's actually really useful. I can just kick these really useful. I can just kick these really useful. I can just kick these agents off and many of these have been agents off and many of these have been agents off and many of these have been running running running On this one, you can't really see it. On this one, you can't really see it. On this one, you can't really see it. Let's see. Let's see. Let's see. Yeah, so this one has been running for Yeah, so this one has been running for Yeah, so this one has been running for 12 minutes autonomously already, and 12 minutes autonomously already, and 12 minutes autonomously already, and before that it was running for 10 before that it was running for 10 before that it was running for 10 minutes, too. Can I see it on this one? minutes, too. Can I see it on this one? minutes, too. Can I see it on this one? No. Okay, so No. Okay, so No. Okay, so Basically, these have already been Basically, these have already been Basically, these have already been running for I don't know, 20, 30 minutes running for I don't know, 20, 30 minutes running for I don't know, 20, 30 minutes straight. So, I can just give them a straight. So, I can just give them a straight. So, I can just give them a very long task, and especially Fable is very long task, and especially Fable is very long task, and especially Fable is really good at this. I give them a very really good at this. I give them a very really good at this. I give them a very long task, and I just go about and do my long task, and I just go about and do my long task, and I just go about and do my other work. And then, if I hear the other work. And then, if I hear the other work. And then, if I hear the bloop sound, then I can bloop sound, then I can bloop sound, then I can go back and review. So, for that, this go back and review. So, for that, this go back and review. So, for that, this is actually super useful. I don't think is actually super useful. I don't think is actually super useful. I don't think you can hear it, but actually, if I type you can hear it, but actually, if I type you can hear it, but actually, if I type clear, it will trigger the sound. clear, it will trigger the sound. clear, it will trigger the sound. No, it doesn't. If I exit this one, No, it doesn't. If I exit this one, No, it doesn't. If I exit this one, and maybe if I start it again, and maybe if I start it again, and maybe if I start it again, No, you won't I I don't think you hear No, you won't I I don't think you hear No, you won't I I don't think you hear it, but that doesn't matter. It plays a it, but that doesn't matter. It plays a it, but that doesn't matter. It plays a sound when I do this.
-
sound when I do this. sound when I do this. So, for that I think this is is very So, for that I think this is is very So, for that I think this is is very useful. useful. useful. Uh I I really like it because it um Uh I I really like it because it um Uh I I really like it because it um helps me manage all of these multiple helps me manage all of these multiple helps me manage all of these multiple agents and agents and agents and uh finally, uh finally, uh finally, this was maybe the biggest hurdle, but I this was maybe the biggest hurdle, but I this was maybe the biggest hurdle, but I am actually running all of these AI am actually running all of these AI am actually running all of these AI agents inside of dev containers. So, I agents inside of dev containers. So, I agents inside of dev containers. So, I never run anything locally. I always run never run anything locally. I always run never run anything locally. I always run anything everything in containers. In anything everything in containers. In anything everything in containers. In this case, I'm I will make a second this case, I'm I will make a second this case, I'm I will make a second video about my actual AI setup, but I'm video about my actual AI setup, but I'm video about my actual AI setup, but I'm using something called Life OS, which I using something called Life OS, which I using something called Life OS, which I always run inside of a uh always run inside of a uh always run inside of a uh dev container. dev container. dev container. And the way I do it is I I set this And the way I do it is I I set this And the way I do it is I I set this environment variable her agent Claude environment variable her agent Claude environment variable her agent Claude dev pod SSH, and then I SSH into my dev pod SSH, and then I SSH into my dev pod SSH, and then I SSH into my container, and then it's going to start container, and then it's going to start container, and then it's going to start another session inside of that another session inside of that another session inside of that container. So, all of these sessions are container. So, all of these sessions are container. So, all of these sessions are actually running inside of the same actually running inside of the same actually running inside of the same container. And that allows me to run container. And that allows me to run container. And that allows me to run Claude with dangerously skip Claude with dangerously skip Claude with dangerously skip permissions, and I don't have to worry permissions, and I don't have to worry permissions, and I don't have to worry about it messing with my about it messing with my about it messing with my actual computer or my file system. And actual computer or my file system. And actual computer or my file system. And as you see, I exit it now, and it as you see, I exit it now, and it as you see, I exit it now, and it immediately disappears from here because immediately disappears from here because immediately disappears from here because I run it in in in containers, I don't I run it in in in containers, I don't I run it in in in containers, I don't have to worry about it messing with my have to worry about it messing with my have to worry about it messing with my file system or any anything important.
-
file system or any anything important. file system or any anything important. It only has access to what I give it It only has access to what I give it It only has access to what I give it access to. access to. access to. And And And uh so, I I managed to figure out how to uh so, I I managed to figure out how to uh so, I I managed to figure out how to do that with Herder and dev containers do that with Herder and dev containers do that with Herder and dev containers as well. I I also believe that Herder as well. I I also believe that Herder as well. I I also believe that Herder has like a phone app coming up. I'm not has like a phone app coming up. I'm not has like a phone app coming up. I'm not sure actually, but if you have Termius, sure actually, but if you have Termius, sure actually, but if you have Termius, like your your normal um terminal on like your your normal um terminal on like your your normal um terminal on your iPhone, you can literally just SSH your iPhone, you can literally just SSH your iPhone, you can literally just SSH to a server and then run Herder from to a server and then run Herder from to a server and then run Herder from there. So, that's always a way to do it. there. So, that's always a way to do it. there. So, that's always a way to do it. But yeah, if you're running But yeah, if you're running But yeah, if you're running multiple AI agents at the same time, I multiple AI agents at the same time, I multiple AI agents at the same time, I actually think this is a good um a good actually think this is a good um a good actually think this is a good um a good solution. I wasn't solution. I wasn't solution. I wasn't I was a bit skeptical in the beginning, I was a bit skeptical in the beginning, I was a bit skeptical in the beginning, but now that I am building my own but now that I am building my own but now that I am building my own applications again and I'm I'm working a applications again and I'm I'm working a applications again and I'm I'm working a lot with multiple AI agents, I now see lot with multiple AI agents, I now see lot with multiple AI agents, I now see that that there is actually a lot of that that there is actually a lot of that that there is actually a lot of value in this one, especially if you're value in this one, especially if you're value in this one, especially if you're command line focused like I am. And if command line focused like I am. And if command line focused like I am. And if you're curious what I'm building is you're curious what I'm building is you're curious what I'm building is Kubecraft AI. This is trained on over Kubecraft AI. This is trained on over Kubecraft AI. This is trained on over 300 hours of coaching 300 hours of coaching 300 hours of coaching coaching calls that I've done by coaching calls that I've done by coaching calls that I've done by mentoring mentoring mentoring aspiring DevOps engineers into aspiring DevOps engineers into aspiring DevOps engineers into six-figure tech careers. So, I teach six-figure tech careers. So, I teach six-figure tech careers. So, I teach people how to land these jobs. I have people how to land these jobs. I have people how to land these jobs. I have mentorship calls with them and I've been mentorship calls with them and I've been mentorship calls with them and I've been doing that for 3 years and I've gathered doing that for 3 years and I've gathered doing that for 3 years and I've gathered all of those calls, all of the over all of those calls, all of the over all of those calls, all of the over 6,000 posts that are made in my 6,000 posts that are made in my 6,000 posts that are made in my community community community and all of that is I I built an AI and all of that is I I built an AI and all of that is I I built an AI system on that. So, for example, I can system on that. So, for example, I can system on that. So, for example, I can ask it, "How long should I study before ask it, "How long should I study before ask it, "How long should I study before taking the CKA?" And then I once
-
taking the CKA?" And then I once taking the CKA?" And then I once mentioned that I used about 80 hours of mentioned that I used about 80 hours of mentioned that I used about 80 hours of focused study for this and then it will focused study for this and then it will focused study for this and then it will resurface that. It will then show me the resurface that. It will then show me the resurface that. It will then show me the sources of those those and my students sources of those those and my students sources of those those and my students can click and open that specific post in can click and open that specific post in can click and open that specific post in the school community. So, it's a really the school community. So, it's a really the school community. So, it's a really useful way of resurfacing older useful way of resurfacing older useful way of resurfacing older information and to help people getting information and to help people getting information and to help people getting support with any with all of their support with any with all of their support with any with all of their career and tech questions as they're career and tech questions as they're career and tech questions as they're going through the Kubecraft curriculum. going through the Kubecraft curriculum. going through the Kubecraft curriculum. And it's working really well. Oh, just And it's working really well. Oh, just And it's working really well. Oh, just this week we've landed almost six jobs, this week we've landed almost six jobs, this week we've landed almost six jobs, I believe. So, Rodolfo, I believe. So, Rodolfo, I believe. So, Rodolfo, David, Erden Bolt, Kevin, David, Erden Bolt, Kevin, David, Erden Bolt, Kevin, Justin, Justin, Justin, and Marco, they all and Robert, they all and Marco, they all and Robert, they all and Marco, they all and Robert, they all landed jobs. landed jobs. landed jobs. So, in in just 1 week. So, yeah, we are So, in in just 1 week. So, yeah, we are So, in in just 1 week. So, yeah, we are really good at what we do and this really good at what we do and this really good at what we do and this application really takes all of the application really takes all of the application really takes all of the information we've built up over the past information we've built up over the past information we've built up over the past 3 years and make that accessible. But I 3 years and make that accessible. But I 3 years and make that accessible. But I this I'm running this myself on this I'm running this myself on this I'm running this myself on Kubernetes and I'm constantly developing Kubernetes and I'm constantly developing Kubernetes and I'm constantly developing it. I'm working on it. This is like the it. I'm working on it. This is like the it. I'm working on it. This is like the the place where I spend my tokens and the place where I spend my tokens and the place where I spend my tokens and I'm having a lot of fun with this. So, I'm having a lot of fun with this. So, I'm having a lot of fun with this. So, it's um this is how how I am currently it's um this is how how I am currently it's um this is how how I am currently um basically this is where my need for a um basically this is where my need for a um basically this is where my need for a a multi-agent system comes from.
-
a multi-agent system comes from. a multi-agent system comes from. Before this, I was just fine with just Before this, I was just fine with just Before this, I was just fine with just one Cloud Code instance running straight one Cloud Code instance running straight one Cloud Code instance running straight on my terminal. So, before you get into on my terminal. So, before you get into on my terminal. So, before you get into this, you really have to think about do this, you really have to think about do this, you really have to think about do I actually really need this? Do I Is I actually really need this? Do I Is I actually really need this? Do I Is this something that I um is going to this something that I um is going to this something that I um is going to improve my workflow or is it just going improve my workflow or is it just going improve my workflow or is it just going to add complexity? But, for me in the to add complexity? But, for me in the to add complexity? But, for me in the case where I'm running my AI agents on case where I'm running my AI agents on case where I'm running my AI agents on separate machines, I have multiple separate machines, I have multiple separate machines, I have multiple projects going on at the same time with projects going on at the same time with projects going on at the same time with multiple long-running tasks. multiple long-running tasks. multiple long-running tasks. Uh and I am command-line focused, so I Uh and I am command-line focused, so I Uh and I am command-line focused, so I want everything on the command line. Uh want everything on the command line. Uh want everything on the command line. Uh in that case, Herder is actually the in that case, Herder is actually the in that case, Herder is actually the tool for the job. So, let me know in the tool for the job. So, let me know in the tool for the job. So, let me know in the comments, do you use Herder or not? And comments, do you use Herder or not? And comments, do you use Herder or not? And if you are interested at becoming a if you are interested at becoming a if you are interested at becoming a DevOps engineer, you can go to DevOps engineer, you can go to DevOps engineer, you can go to roadmap.cubecloud.dev. roadmap.cubecloud.dev. roadmap.cubecloud.dev. There will also be a link in the There will also be a link in the There will also be a link in the description below. You can fill out the description below. You can fill out the description below. You can fill out the the questionnaire. It will put you on a the questionnaire. It will put you on a the questionnaire. It will put you on a certain stage on the roadmap and it will certain stage on the roadmap and it will certain stage on the roadmap and it will send the free roadmap to you. It's a send the free roadmap to you. It's a send the free roadmap to you. It's a 45-page document that's going to tell 45-page document that's going to tell 45-page document that's going to tell you exactly where you are in your you exactly where you are in your you exactly where you are in your journey and what stage you are in and journey and what stage you are in and journey and what stage you are in and how you should develop from that stage.
-
how you should develop from that stage. how you should develop from that stage. So, it tells you exactly what you need So, it tells you exactly what you need So, it tells you exactly what you need to do to get closer to landing that to do to get closer to landing that to do to get closer to landing that six-figure DevOps job. It's completely six-figure DevOps job. It's completely six-figure DevOps job. It's completely for free. The link is in the for free. The link is in the for free. The link is in the description. Check it out and I will see description. Check it out and I will see description. Check it out and I will see you in the next video.
Summary
The main theme is an analysis of Herder, a new tool for managing AI agents. Key subjects mentioned include GitHub stars, recent funding, and a comparison to tmux for AI agent status tracking. The practical takeaway is that Herder offers a dedicated and potentially more reliable solution for managing AI agents, especially compared to custom tmux configurations, and may evolve into a SaaS offering.