Codex is FREE Forever...Here's how!

**Julian Goldie** (0:00)
Codex is free forever, here's how. So what we have over here is Codex, and this is still the same CLI, but the good news is that you never need to pay for it again because there's a free model that you can plug into it called Omniroute. And Omniroute allows you to use over 93 providers and code with Codex directly. And we can also plug in other APIs, for example, like HY3, and use this completely free. Now, if you're wondering, okay, how does it work, or what does it look like, or you can create with it, etc. Here's some examples, so we actually have a workspace here where we can build with this. And you can see, for example, Codex builds free via Omniroute. That's something that we created inside the workspace. You can create nice stuff, as you can see right here. It's pretty simple and easy to do. You can build out games, websites, apps, tools for free. And this is a powerful system when you use the power of OpenAI's Codex with a free brain inside it, so you still get the same power of it. So this is something I call the free Codex engine. It's OpenAI's coding agent running on OmniRoot, and you can also plug in Tencent's HY3, which is another free API that's available until 21st of July. And then after that, you can use OmniRoot, which is free forever anyway. So what we actually did is we took Codex and we plugged it into our agent operating system. The reason that we did that is because it links to our memory. We've got all our tools in one place. It's easy to use, et cetera. And we can easily switch to Codex over here whenever we want. And we can code with it inside the chat. We can also switch between auto-approve, ask, yellow mode. And then we have the goal mode as well. So we can have it running autonomously for hours for free using this system as well. And so Codex is OpenAI's coding agent. It can plan, it can write, it can run commands. Normally it runs on OpenAI's paid models, but Codex has a hidden superpower, which is you can point to any model with one configuration file. So I pointed it at HY3, which is Tencent's new 295 billion parameter, which is free on OpenRita right now. And I dropped it into the Codex tab of our system. So we have three pieces to this. Codex is a hand, HY3 is a brain, and the bill is still zero. So there's three pieces. Codex is a hand, so OpenAI's coding agent. You give it a task, it writes the files, it runs the commands for reports back. The agent itself is free to install. You would normally pay for the model behind it, but in this situation, you don't need to. And then the brain. So you can choose between Omniroute, and if you wonder, okay, what is Omniroute? So this is a free API you can get, and it runs locally, so it gives you a local endpoint, and then you point that endpoint into Codex, and you can code with it for free. And you can also use this across all your favorite AI tools, so for example, Claude, GBT. We've actually got a version inside the Agent OS for free Claude code as well, and then we can use two of the most powerful harnesses in the world for coding with a free model. Pretty insane, right? And then we can swap it over. So the brain is the model, Codex is the hands, and then there's the swap, the one config file. So Codex actually lets you define a custom model provider. So one small file tells it, talk to OpenRouter or talk to OmniRoute instead of OpenAI and use that instead. And that's the entire trick, which is that your OpenAI login stays untouched as well. Pretty powerful stuff. Now, also when you set this up, you basically edit the Codex latency here. Now, if you want to get my full system for doing this, like if you want to get it already wired in to one agent operating system, you get that in the AI Profit Boardroom. Think in the comments description or just go to the aiprofitboardroom.com and you can get the Agent OS system inside there, ready to download and install. So just go to the AI Profit Boardroom over here, then you shall go to the classroom. And then inside the new daily update section, you can get the Agent OS system. As you can see right here, we actually update it daily. So it just got updated today. You get a video tutorial, a full step-by-step guide on how to use it. And you get the zip file as well as all of our best trainings inside our AI community. Now coming back to the free Codex engine. So when you're using this, you can plug in either if you're using HY3, you would plug in your API key for that, or you can use your OmniRoot as well. And then Codex packs the task with its tools. So you're basically your prompt goes out with the tool list. You know, I can write files, it can run commands, it can read files. But the difference is that the brain is what's acting, which is the API. Now also you might say, okay, why would you use Codex to do this? Well, the thing is, Codex is a great coding agent. So you can still get the most of their agent harness, but you don't need to pay for the API. So you get the best of both worlds. And if you're wondering how this works, so you got your task, either in the tab or the terminal, you got your profile loading, OpenRooter starts running, HY3 Axe or OmniRoot Axe, and then the files are done. And you know, the difference here is like, Codex used to be my favorite pair of hands really. And every task it ran burned paid tokens, because you're using either the API or using your subscription with ChatGPT. And then it can run through tokens really quickly and run out very quickly. And then also, if you're using long agent loops, like we have a loop engineering system inside our system here with the goal, which means like if you give it a goal and you let it run overnight, it can do amazing things, but it can also run out of tokens very quickly. So what happens in the middle of the night when you run out of tokens? Well, you can use OmniRoot or HY3 and then you don't need to worry about it. So free models before as well, like they either couldn't drive an agent or they fell over on tool calls as well halfway through. So the expensive brain did everything, including jobs, a free one could handle. Then because we've got free APIs on OpenRoot and we've got OmniRoot as well, Codex only needs one file to switch your brains. So now Codex can run on free models that cost nothing. It can one-shot complete landing pages or tools, apps, or games for it. We can switch the engine. So if we want to switch back, no problem. We just go inside the chat here and we would switch back to the one we want. If HY3 gets rate limited because it's a free API, then we can go to Omniroot. If Omniroot isn't what we need right now, then we can switch to GPT 5.6. It's so easy to switch between each one just with a little drop down. Also, that is way easier than if you were inside, for example, ChatGPT trying to change your model here. This is so much work and you won't get that drop down for the free APIs, you can't customize this system in the same way. Whereas if you're using the CLI for Codex, which is what we're doing over here, way easier, way more chill, way easier to do. Now you also might say, okay, aren't like agent operating systems and that sort of thing, pretty difficult to run. We actually have over 198 pages of people absolutely loving the agent operating system and you don't need to be technical to do it. Like I'm non-technical, I don't understand code, I'm not a programmer or anything like that. But if I can do it and if everyone here can do it inside the community, then you can do it too. You don't need to be like an absolute AI whiz kid to do this stuff. So here's an example of a landing page that we built. We can create pretty nice stuff with this. We literally just said like build out a beautiful landing page and actually created it and saved it inside a workspace. That's the other thing as well. It's like if you're using Codex directly in the Chat Cheapity app, so the Chat Cheapity app just rebranded to Chat Cheapity work instead of Codex. But if you have a look here, you can see that you can lose everything you've created very easily. You can't preview everything that you've built. It's very difficult to see what you've created. And also it's very easy to lose a project. So how do you get around that? Well, what you can do instead is you have this workspace over here. So you can save everything and come back to whatever you want whenever you want. And it's easy to do. We actually ran Goal Mode on this as well. We said create a website in a game and on the website showcase the game. Pretty simple, but it just ran in Goal Mode creating an awesome game. And then you can see the example of the website right here. So it just looped around and got that chopped up, which is pretty cool. Now, some people will say, is it technical to wire up? If you do this yourself a little bit. But I mean, if you use our agent operating system instead from the AirProf one, it's going to be way easier. And then the free Codex engine is pretty simple because it roots out to the models once you set them up. So you only set the models up once. You also might think, okay, does an agent operating system use a lot of tokens? But you can just plug free models into all of this system. And then the old way is like every task was metered, right? So Codex runs on paid models only before each loop costs and gets expensive. You ration experiments because the meter is always on. Free models can't hold an agent loop, so you never try looping or anything like that. You test a new model and that means new keys, new tools, new billing, and then things get really expensive and you just don't use it at all anymore. With this system, with the free Codex engine that I'm showing you today, you have one profile file, then the engine drop down to choose the model. You get real agentic builds that we've tested, and you can get a 260k context window agent for free. And then you get your whole OpenAI setup untouched, so it doesn't mess around with the app locally. If you're using this, you don't need to worry about what you do inside the AgentOS. It's totally separate, which is nice as well. So if you want this whole system with the free Codex engine setup, you can get that inside the AirProper volume. Link in the comment subscription and go to theairpropervolume.com. This whole agent operating system is amazing. We build cool stuff like this inside there all the time. You can actually get the Omniroot Coder over here as well. You get free Clawed code, you get HY3 Coder, memory system, and all sorts of amazing tools. And we're just constantly improving this every single day as well, based on the community feedback. So feel free to get that link in the comments description or go to theairpowerforbomb.com. Inside the community, you can ask questions, get help and support in real time. Inside the classroom, you can get all of my best trainings. If you want the Agent OS system with the free Codex engine, you can get that over here. If you want to learn more about Omniroot, you can get that over here. And if you want to reduce the amount of tokens you use in general, you can check out the token minimization. Playbooks, we'll just have all of our best trainings inside the classroom. In the calendar, you can drop a weekly coaching course, ask questions, get help and support in real time. Inside the map, you can meet people locally who are building with AI like you. And that's all available inside the AI Profit Boardroom. Link in the comments description or go to the aiprofitboardroom.com.