Stop paying for AI coding tools. Here's what I use instead

追加: 2026-05-10

[00:00:00]You are paying a monthly fee right now to rent access to an AI that runs on someone else's server, processing your code on someone else's hardware and sending your intellectual property through someone else's pipes. And the moment you stop paying, it's gone. Now, here's the uncomfortable question nobody in the AI space wants you to ask. Why are developers, people who are literally building infrastructure, the ones who agreed to this deal? See, the average developer spends over $1,200 per year per developer just to write code faster.

[00:00:28]Now, Cursor alone saw pricing complaints spike so hard they made headlines.

[00:00:32]GitHub Copilot Enterprise tier hit over $100 per month, and OpenAI's usage based API billing like developers are waking up to four figure invoices on projects that they thought were small. Now, meanwhile, your machine is sitting right there behind you doing nothing with GPU power enough to run a local model that doesn't care about your token count.

[00:00:52]What if the subscription was never necessary? Let's dive into this today.

[00:01:05]Welcome to Startupac. I'm Spencer and here at Startupac, we love to build custom software solutions for companies.

[00:01:10]With a decade of executive leadership as a fractional CTO and 25 years in software development, I have helped transform tech teams and products, including building out custom AI solutions. Now, we just talked about the money, and it's real. But this isn't just a cost conversation. This is a control conversation. Who owns your AI workflow? who owns the context of your codebase when it goes through a thirdparty API and who gets to flip the switch when the pricing model changes again. So today I'm going to do a full live demo of open monoagent.ai, an open source terminal native AI coding agent that we built from scratch that runs on local LLMs cost you absolutely nothing to use because we gave it to you for free and it installs with single command. Now I'm going to break it down.

[00:01:49]I'm going to show it all to you with my own hands. We're going to walk through the whole thing as a demo today. Now, biggest compliment you can give me is to drop a comment and also to go out to the GitHub and leave a star on this because we're going to be walking through this today. And I really want to go through and show you guys this today because I'm really excited to be able to demo this.

[00:02:06]Now, before I dive in and demo it, the SAS subscription model works great for business apps. It falls apart when the app is generating the code that touches your production infrastructure. Think about what you're actually sending.

[00:02:18]proprietary business logic, unreleased features, internal architecture, all of it going through a third-party server every time you hit tab complete. So, I've spent a long time in software development and it's kind of crazy that how we've lost our mind and we're sending all of our secrets up into the cloud to give it away to give all this data out to people to these big frontier models. The biggest cloud providers built enterprises on promise of just pay for what you use and what developers are using is adding up really fast. Uber drove through their entire budget in under four months for the whole year.

[00:02:50]Per token billing is particularly brutal for coding workflows because code completion is chatty. You're generating tokens constantly. And we're going to show you this, right? The best part again is about the control though.

[00:03:01]There's a moment every technology curve where the good enough line it gets crossed. And for local LMS running on consumer hardware, that line is behind us now. There's a lot of great open source models and we've tweaked these and tuned them and built some framework and an agent that is open source and we're giving away to you for free. So I started with open mono agent because it's a real refactoring task and sorry I started this using this for a lot of our production projects. The gap between local and cloud it like it's definitely narrowing. Now I'm not going to say that we're going to have every single feature that cloud code's going to have but I can tell you that for most people this is going to be good enough. Every popular AI coding tool eventually adds an IDE plugin and we're working on our own right now, but right now we started with just terminal native tools that sit at the layer that developers actually trust. The same layer where git lives, where your build tools live, and where your CL uh where your CI pipeline lives.

[00:03:55]So, Open Mono Agent is a single command.

[00:03:57]Install it, point it, point it uh at the service. And I'm going to show you how to connect these here. It's super easy to set it up and I'm going to show you how to do it. We walk you through all of it. So when your AI coding agent is closed source, you have no idea what it's doing in the background, what it's sitting there doing with your code. But in this case, we give you the full open- source license for open mono agent. And that means the code is auditable, forkable, and nobody's going to quietly push an update that changes the data handling policy. So I've watched the AI tooling space move goalposts on privacy terms twice in the last 18 months. Once is a mistake, twice as a pattern. Open source also means the community can extend it. You can build on this to be whatever you need it to be. Now, we've got some great features and I'm not going to get to go through all the features today because I want to give you a demo and show you how easy this is to get installed because for free you're going to be able to download and open this. We're also giving away a free inference box that I'm going to show you how to sign up for here at the end. So, make sure you stay to the end of the video. Now, without any further ado, I want to jump in and get started on this.

[00:04:58]Okay. So, AI shouldn't have a meter.

[00:05:01]Unlimited tokens forever. Yep, forever.

[00:05:04]your machine, your agent, use it from anywhere. Open Mono Agent AI is a terminal native coding agent powered by local LLMs, 100% open source, free forever because we're giving away to you guys and installed with a single command. We proudly build it on top of C.NET because AI tooling should be infrastructure, not a subscription. And here's our manifesto. I'm really proud of this part. AI shouldn't be a subscription you rent. It should be infrastructure you own, sitting on your desk, like back behind me here, serving your code, answering only to you. Now, I'm going to show you and it's got great tons of great features. Now, the box that I'm going to be showing you on here is right where our sweet spot is, right?

[00:05:41]It's running RTX3090. This is going to run really, really fast. These bigger cards are great, and we're actually doing a giveaway of this little box right here that I'm going to show you how to sign up for here at the end. And this will give you about 20 tokens per second, which is good enough. If it's not blazing fast, the 3090 is definitely the sweet spot. So, you definitely want to try to get your hands on one of those. And that's what I'm going to be demoing for you guys today. Now, what we're going to do is I've already logged into this box. And what I'm going to do is I'm just going to come up here. I'm going to copy this. You can see it just copied here. And I'm going to paste it in here and let it fly. Okay. It's going to pull it down. It says, "Hey, do you want to install both?" Now, you could install both, but you'd be running the terminal and all your projects and everything would be on this. I'm going to take treat this as my infr server. So I'm going to pick number two on the server. Now this is um going to run on this infr server here. I don't have anything else installed on this. So this was a brand new YUbuntu 2604 install.

[00:06:36]You can see it's going to download all the different prerequisites that you need. Now part of these prerequisites may take a little bit of time to download an install. Kind of depends on what your bandwidth is. Now, while while this is going through is here, I'm going to continue to go through some of the features that we have here because I want to be able to show you guys some of the awesome stuff that we're building out here. So, while this is going here, we're just going to keep cruising through. Now, you notice that it detected that I have my GPU. It's actually going to even install all the drivers that you need for it. Right? So, it says, "Hey, you're going to need to do a reboot after this." And that's to be expected. All right. Now, while this is installing here over there, we're going to actually go through some of the different features we have here cuz this is a full coding agent that lives in your box. So, this is not uh something that anybody is going to need to install. Now, I'm going to pause this for just a second while this install runs through. And while this is installing here, I want to show you back behind me. So, this is a box running here behind me. I got this box for about $500 off of Facebook Marketplace. I'm sure you can go out to your Facebook marketplace, find a similar type box. I put a 3090 into it. All in, I'm a little over $1,000 in on this box. Okay, now that's a little high and a lot of people are like, "Man, I can't spend that much." One of the great things is we're going to do a giveaway on one of these type of boxes where you can run the local imprints yourself and run it on on the model here. Now, um a as you run through this, it's pretty simple. So, there's not a lot on this box. It's not a super powerful box. the actual mach, the motherboard, the processor, I usually like to find at least to have 32 gigs of RAM because it finds the model uh swaps in and out faster. But other than that, really the GPU is the sweet spot here of what's installed on there.

[00:08:11]All right, so you can see that it ran through and installed the drivers here, right? Installed the drivers that we needed and then it said, "Hey, would you like to reboot?" So this case, I'm going to tell it yes. We're going to go ahead and let it reboot. It's going to reboot and you're going to have to log back into that. Okay, so uh I'm going to let it reboot. It's going to take a few seconds here. While we're going through the reboot here, you can see some of the different features that we have here, right? So, we can see that uh you know we have embedded inference, zero setup, right? Like you can see how simple the setup was one command, right? 2 built for long sessions, Docker sandboxed, 20 plus MCP tools, right? Built for.NET focused on.NET. We have LSP for C and TypeScript. We have playbooks. This is our version of skills, but this is much much better. And I'm going to demo a lot more of these playbooks for you in the coming weeks. We have this dual box mode which I'm going to show you. This is pretty awesome because you can we actually set up our own relay server so that you can actually connect to your inference box from anywhere. So your agent can follow with you on your laptop. Everything else connects here.

[00:09:08]Let me get connected back up to the server here and finish the install to show you guys here. Now, one of the things I do want to point out to you is right before it rebooted, it says, "Hey, after the reboot, make sure you run this open mono setup." So it tells you exactly where to resume your setup here.

[00:09:20]So, we're going to grab this. We're going to rerun it here. We're going to say, "Hey, we want to do the inference again, right?" Because it had to install those drivers. So, see if I can type my password correctly. This isn't a super secret password because it's not really a production box. This is just what I'm using for my inference, just for the GPU. So, no data ever actually lives on this box. It's just the workhorse. Now, once it, see, it can detect the Nvidia drivers already installed. Now, it's going to install some of the Nvidia CUDA toolkits. So, we're going to go back, keep going back through here. While this guy's running here, I'm going to go through some more things. So, we also have persistent sessions. This is a very unique thing that we're doing where you can do checkpoints. Kind of like if you're playing that game and you want to save those checkpoints to come back to later. Now, like I talked about, we actually do have this and we are doing some work to support smaller models.

[00:10:06]Now, I'm going to throw a big caveat right here because the sweet spot is the 3090. You really don't want to go a whole lot more than smaller than these models here. But if that's all you have is a smaller one, you can just know your accuracy is going to be less. These are the best ones to run your topline coding models on. Now, we can go through and talk about how it stacks up. We can talk about a lot of different things, but ultimately this one command gets you started and you're off and running. So, I'm going to pause this here for just a minute while it runs through and installs this toolkit because the CUDA toolkit can take a little bit to go through. All right, you can see it's still chugging away here. Installing the Nvidia container toolkit, adding Docker, you know, installing some of the Docker packages. All of these are going to be important for our build. Installs.NET 10 because that's what we're built on here. And you, trust me, you'll thank me later. I know get a lot of people are coming back. Now, uh, you know, this is then actually going to go where it's going to go and download the model.

[00:10:57]Okay. Now, this is definitely going to take a little while depending on your bandwidth. You know, I've got a decent bandwidth here, but for a model that's, you know, 15 gigs here, uh, this is going to take us a minute. So you can see that it ran through and did all the prerequisites, set up a lot of different things, you know, set up my directory, um, and then ran this. So again, I'm going to pause while this downloads cuz like we don't want to watch paint dry here. But it's going to actually download this model here for you and then we'll show you here at the end.

[00:11:21]Now, while that is downloading, I do actually want to show you something that's pretty exciting that's a new uh app at edition here. We've actually just launched our own agent where you can actually connect to your own inference and have your own chat going against your own inference. Now, we're going to continue to build out more controls here where you can actually control the agent yourself and there's going to be a lot of cool things we're building out here in the future. We've got tons of great ideas, but we're actually So, this is so you can hit this whether you want Apple or or Android. Um, this is uh this model works great. Um, I can show you some of this in a future demo here, but it works awesome and I'm going to show you how you can set that up with your inference.

[00:11:58]Again, free. Go download the app. It's free as in free. Now, while it's also downloading, I want to go and show you the uh repository here. Now, as always, you know, if you can do me a big favor and, you know, hit that star there because, you know, whoops, I guess I'm not signed in right now, but if you hit that star there, that would be awesome.

[00:12:14]But this is the repository that we have here while this is downloading. I'm going to run through a little bit of this for you. So, you can see that we have the same, you know, single line install script. We're going to add a little bit more here about how to set up the tunnel and some of those other things, but we've got some different pieces. Goes through, you know, what's exactly inside of it. talks about how the coding agent which different models we're using but we go through all of this but most importantly here you can see we've got tons of different documentations so definitely dig into the documentations read up on playbooks right playbooks is going to be a huge thing and like I said we're going to talk more about this in the future here but also if you want to see what the architecture is this goes through the top level architecture of how we're actually working with the code with the agent uh the breakdown of all of it I'm not going to go through this but this will go through actually all of the different pieces of how exactly we built this. Um, one of the other parts too is that we're going to talk about the code review graph, right? So, this will actually allows you to load up template projects into your uh into your own local agent and it will work against these template projects. It's kind of think of it kind of like rag, but it's a little bit different. This actually will allow will help to keep the model in line with what you want for your coding standards. So, it's pretty exciting.

[00:13:23]This is an awesome opportunity. Now, you can also do the same thing with Graphify. This is kind of like having a PM built, right? A product manager built it right into your agent itself. So, lots of great opportunities, lots of great tools that we have here, but we're going to keep cruising here. Okay, looks like my model finished downloading.

[00:13:40]We've got the Nvidia GPU detected. It's going to stop any of the containers.

[00:13:44]It's starting to run Llama CP, the Llama server, right? And so, this is what we're using to actually host the GGUF.

[00:13:51]Um, and so you can see this is starting to start up and we're going to check on this in a second after it's got this up and healthy. All right. And there you go. You see it finished. So it says, "Hey, your llama server is healthy." The port that it's running on, here's the model says it's being running in GPU mode. Now, one of the things that I definitely want to point out here is don't use this link here, but grab this open model tunnel setup. This is a really important part here. Now, I'm going to kind of step through some of this, uh, and I'm going to definitely run through, you know, some pieces of this. Um, and so we're going to give you kind of some, and I'm going to kind of cooking show this here because I don't want to show, you know, everything here.

[00:14:23]But what you're going to do is you're going to actually, so on the inference box, you're going to run this tunnel setup. This is actually where you're going to enter your email address. Now, I'm going to pause this here for a minute because again, I don't really want to get a billion spam messages. So, I'm going to run through this really quick on mine. All right, so I typed in my email address and I got this email here. Now, this is pretty cool here because this is the email address that we that you or sorry, this is the email and you're going to get a code here.

[00:14:48]Once you get this code, you pop this code in here. Now, this is going to actually say verifying your information.

[00:14:54]And from here, it's going to actually do some stuff and it's going to give you a set of configurations. Now, these configurations are also going to get emailed to you. And I'll show you the email here that they come in. But you want to save this. Now, I'm going to clear this afterward because I don't want you all connecting to mine, but like this for right now is going to be live. So, if you were to go grab these set of configs, which I know you can see on the screen here, right? But I'm going to clear these out here. If I go grab these set of these configs, then you can run open mono agent on your local box.

[00:15:22]And that's what we're going to set up here on our local imprints. And I'm going to show you exactly what that looks like next. Okay, but you're going to want these configs to be able to connect your agent to the inference server, right? That's what we're trying to do here is this is going to allow us to connect that agent over to the infr server no matter where they are, no matter what network they are. And we're going to host that relay server for free. We're not doing anything. We don't actually see the traffic. This is just to make the peer-to-peer connection first and then from there we have we have we're nowhere in the mix. So this is totally secure completely encrypted end to end. So let me jump over and show you this part here and how this is going to work. So I'm going to come over here to the website. We're going to copy this command. We're going to run it over here. See now what when it comes and prompts me after I've run this. You can notice that. So I I paste it in the command here. It's running. It pulls it down. I it asks then do you want to install this on this machine? I say yes or I say I don't want to install both.

[00:16:14]In this case, I want to install just the agent only. Right? So this agent only is all we want to install here. So it's going to run through the open monop prerequisites runs. Once it checks all the prerequisites, it asks if I want to install with the GPU. In this case, I don't because I'm just installing the agent here. So again, the agent, as you can see here, pretend this is the agent.

[00:16:34]This is actually the mobile app, but it's a similar graphic. The the agent actually is going to run against the inference. Now, this is the secure relay that we're going to set up here in a second. We set it up over on the infrance side. We're going to connect to it on the agent side. So, it runs through. It downloads everything. And now, once it gets to the bottom here, it's actually going to come in and say, hey, it checks for the net. Make sure we have all the everything needed. Now, it says, hey, your installation is complete for the RO agent. Right now, one of the first things we need to do is we have to actually reload uh the command here so that it can actually reload the script so that we can use this. So, let's reload bash. Now, from here, we're actually ready to rock and roll. Okay, so we have our open mono agent. We can take a look at it and say, hey, what's all the different commands that we have here? So, let's uh run our open mono agent with the help. These are all the different commands that we have. Now, one of the things we definitely need to do is we want to get our tunnel set up because we need to connect over to um and you can see here it says point this agent at your inference server, right?

[00:17:34]So, we got to go get our commands, our configs here. So, we're going to go back to that email that we had open, right?

[00:17:40]So, we have this email open here. So, we're going to copy this one. This is the first line. So, first thing we're going to do is copy that one in here.

[00:17:48]Sets that. We're going to copy then our API key. And I know everybody's going to go use my API key. Don't worry, by the time you see this video, this will be purged. Um, now once we have this, okay, so we have our a we have this in the agents now set. So, our configs are set.

[00:18:05]I'm going to actually build out a site.

[00:18:06]So, we're going to we're going to take our first test run on this here because I'm going to show you guys real and we're going to go in here and from here we're going to say uh in an actually now we're going to switch over. All right, this is like a little bit crazy large here, but we're going to roll with this here. So, I'm going to say open mono agent. Okay, now it's pulling all this up. So, let's take a look here. And so, you can see that everything's loaded.

[00:18:30]You can see our inference counters here.

[00:18:32]Now, we're going to create a site here.

[00:18:34]So, uh, I'm going to kind of push this keyboard over here a little bit and say, uh, build me a website using React.js.

[00:18:49]Make it all about how amazing program uh, building websites is.

[00:19:00]talk about amazing um amazing chances to build awesome software.

[00:19:15]Okay, now it's going to start to connect and you can see it's starting to spin here because it's actually starting to think. So it says, "Hey, I'm going to build a React website empty workspace."

[00:19:24]Okay, so it's going to ask us for some permissions here, right? So I'm going to allow to allow all because this is actually sandboxed inside of a of a Docker container where it's doing all of this. So any of these commands that it's running is actually running inside of the Docker container. So it's not actually running these on my native machine. So it's trying to do a check and you can see it running through and trying to get a bunch of stuff here to get ready for this. Um and so over here we're going to see in a minute it will start producing the files. Um, now again, I'm trying to do this sideways and looking through four different monitors here. But you can see it's gonna actually try to install Node.js because this is inside of Docker. So, it's got to set up all the comm all the scaffolding and everything that we need.

[00:20:01]Got to go at bite everything so that we can get this all built out to start building our website. So, we'll let this guy run here for just a minute and then we're going to let it catch up. But one of the nice things here is you can see we've already burned 7,000 tokens and cruising through and not even really like I'm not paying any attention to how many tokens cuz I don't care. just run on this machine back here behind me.

[00:20:20]Okay, so we're running real here because this is real raw, right? So, it ran into some problems here. Couldn't quite get some things to work here, right? The command timed out after a long time trying to get some things to work with Vite. So, it says, "Look, forget it. I'm just going to actually scaffold the project myself." So, I'm going to go and tell it to do this. You can see it starts to scaffold this all by itself, right? And so, I can scroll up and down through this. Again, all of this is free, guys. Like, I haven't This isn't using claw. This isn't using anything else. This is just using this inference server back behind me. You want to see here? You can see it's running at 97%.

[00:20:50]You can see that I'm doing my Nvidia, right? And you can see that it's running, you know, full 450 watts bouncing back and forth between this.

[00:20:58]So, this machine back behind me is the one running the 3090 Ti. And that's what's actually running the inference here. So, you can see both parts of this right now in real time. And you see it's doing its thing. So, what do we have here? What are the different parts?

[00:21:10]While this is working here, I want to show you. So you can actually see that it's starting to build out the website right here's actually the code that it's starting to build out. Um and it'll do like this is all working in real time guys. Like I'm running real raw right now. So like see the great part about this is that what you have at the end of the day is you have the agent that runs locally and let's let's jump over to the repository here because I want to kind of break down the different parts for you. So see what we gave you is in the setup we're giving these great setup uh that gives you all of the different scripts and everything to do all the different setups. So, this is what we've done and we've optimized this setup here for you to make it trivially simple. But then on top of that in here, we have the actual CLI. We have tests uh and this is the actual CLI that's running here that you see. This is an extensive project.

[00:21:55]We've been working on this for a long time. You won't see all the history back because we've been working on it for weeks and we just dropped it public about a week ago. Um, but this is actually then this the agent that you see running here. And we have a lot more great features coming like we're be building actually a native Visual Studio VS Code plugin here. But in the meantime, you can just use this terminal uh application and it works really well.

[00:22:15]So you can see that this thing's cruising here. Now, so far we've been burning tokens like crazy and we don't care. It's like that's the best part about this, right? So it's telling us that you can see, hey, it just said it was built with React with love, right?

[00:22:26]So all files written now. Let me install dependencies and build. So it's going to actually do the build for us. Make sure that everything's rocking and rolling, right? And again, I don't really care.

[00:22:34]I'm burning tokens as fast as I can.

[00:22:37]That's the best part about this. So with it going, it says, "Okay, clean was built." Uh, sorry, build cleanly. Let's verify. Okay, everything's built. Tells me exactly what it built down. Talks about everything that it, you know, built out here, right? So to get this started, we can start with dev. And so let's let's try this out here. Let's get our our website fired up here and see how it did for us. Okay. All right. So the agent told me to to run npm install.

[00:23:01]So I did the npm install and then I'm running npm rundev. Okay. So then it starts up my website here. So, let's take a look at what my website looks like. Hey, it built me a website about is absolutely amazing. Discover why why building websites is so great. Now, this was a terrible site because I was literally giving it the worst prompt you can possibly give in history. But it just built all of this for me in just minutes. And this is all on a stack that I own. Now, the best part of this here is this is actually all my code written in my spot written here. So, one of the things now I can do then is I can say initialize git repository. Okay, so we're gonna actually have it do initialize the git repository for us. Uh, let's get this guy rolling here while this does this.

[00:23:45]So, hey, need to get some permissions here. So, we can see that it does that.

[00:23:49]Hey, look, now we've got a git repo, right? So, it's actually going to do that. So, says, hey, the git repository initialized here. So, let's take a look here. Uh, so looks like Visual Studio hasn't seen that yet. But you can actually see that if I jump over here to so let's go and kill that get status.

[00:24:10]And you can see that it oh wants me to set up some of my global configs. So that's probably why it's getting a little cranky. But you can see that it actually did initialize the git repository for for me. So again, very powerful, real working tool, guys. I mean, this is pretty incredible. like I'm giving all of this to you guys totally for free. Now you're like, "What's in it for you?" Uh what is in it for me? So let let's kind of talk a little bit about what it is and why it is that Startup Pack did this for you.

[00:24:36]See, Startup loves to build custom software solutions for companies. But one of the things that we found is we found that we're building out a lot of AI solutions for people. We found that we were rinse and repeating a lot of these different pieces over and over again. We also found that a lot of our clients didn't like us pumping any of their code up into the cloud. It always made him nervous to be like you're using cloud code like why where is that data going? What if my code is proprietary?

[00:24:59]Like and came up with a lot of questions. So that's why we started to use uh and we decided to build out our own coding agent. The other part about this is we just wanted to be able to make sure that we were doing this in a cost-effective way. See, we've built out this inference and we do a lot of different custom software with it. We wanted to be able to show everybody what we're building here. So, we're really excited by the opportunity to be able to show everybody what we've built out here with Open Mono because we really believe that we're going to continue to extend this. I have a team that's working on this full-time here, you guys. And this is pretty incredible cuz we're really, really excited by where Open Mono is going to be going because this is a great opportunity and you guys can do this. Now, I promised that at the end of the video I would tell you guys exactly what we're doing, but we are doing a free giveaway. One of these little inference boxes, right? So, what you saw me doing here was running at about 40 tokens per second. This little box runs at about 20 tokens per sec per second.

[00:25:51]So, it's a little bit slower than what you guys were seeing here. Very, very functional though. This still is here.

[00:25:56]So, you want to make sure you go and enter the giveaway because on May 15th, we're going to be doing a free drawing.

[00:26:01]Free as in free. No strings attached. If you want to come on to the um onto the uh to the live here, we're going to do the drawing and then we'll set up another time where you can jump in. But we're gonna actually be giving a free giveaway of this box. So just go to open monoagent.ai. Hit giveaway up here and you're good to go. So you can see our setup instructions. We've got the view source. This is free as in free guys here because we're really excited about this because AI shouldn't have a meter.

[00:26:27]You should be able to use your unlimited tokens forever. We want to be able to de democratize AI. We want everybody to be able to use it. And we're excited about this. So much so that we're going to be giving away biggest favor that I can ask. Make sure you go leave a like or make sure you like and subscribe and also make sure you go leave a star on GitHub. This shows people how quickly and excited they are about this project.

[00:26:48]So, we're going to continue to invest resources into this. I'm very dedicated to grow this into a real platform. Now, curious to hear what you guys think. Do you agree? Do you disagree? love to have a great conversation and if we can help build some custom software solutions for your company, check out startup.com and otherwise we will catch