Codex is FREE Forever, Here Is How...

**Julian Goldie** (0:00)
Codex is free forever, here's how. So what we have over here is Codex plugged into our agent operating system, and then we can plug any free brain directly into it. So we can still use GPT 5.6 if we really want to, but for example, we could use a free API, like for example, Omniru inside ChatGPT Codex, and then that way, we can get one of the best harnesses with a free API plugged into it. And I'll show you how that works in a second. And then also the great thing about this is we have our workspace here, so anything that we create, we can come back to directly. And then all the powers, all the magic of Codex is directly inside this system. The other thing that I like about this is it's fully customizable. So for example, if I go into the normal Codex app, it's kind of messy, it's difficult to see things, it's difficult to organize. Whereas with this system, it's all clean. We've got our chat here, we've got our goal mode over here, we can see our history, we can see our previous sessions, we can see our workspace, and it's just much easier to come back and keep everything organized. And also, it works alongside, for example, our memory systems or code code or anything else we want to use. Let me show you a side by side comparison over here. So if we go into ChatGBT and the normal sort of Codex app, this is what it normally looks like, right? And the problem with this is like, where's all your creations? Where's your workspace? Where's everything you built? You can't find it. And then also, okay, what happens if you run out of tokens? Well, you can't plug in the custom model into this normal app, right? It's pretty difficult. And then also, you know, like there's different chats and there's different history of conversations that will be difficult to find over here. However, if we have a system like this, well, we can easily switch the brain like so. We can easily see everything that was created over here. We have all of our sessions, we have the goal mode and the chat. So now you're probably wondering, okay, where do you get Omniroot? How do you set this up? So you can see Omniroot over here on GitHub, and this basically gives you like 90 free APIs, 90 free models that you can plug into this system. It also works not just with Codex, but also Cloud, Chat, GPT, OpenAI, whatever you want, right? All sorts of different harnesses here. The other thing that has is RTK and Caveman, which are two different token minimization systems, which means that you can basically get more tokens out of the free models as well, so that you don't get rate limited and you never stop coding, which is great because then you can code on free models without having to worry about getting rate limited or having to stop or switch models in between. The other great thing about this as well is it automatically switches to a different model, so if one of those free models fails, no problem, you can switch to another one instead. So let me show you an example of this. Let's say, for example, we like, build out a beautiful SEO website, and then you can see here the Codex is going to start working its magic, and that's running in the background. The other great thing about this as well is we can go off and we can use, for example, like Hermes Agent, which is one click away inside our system here, or we could use Claude, or we could use all of them together, and it doesn't disrupt what we're currently doing inside Codex. Plus, with OmniRoot, for example, you could plug this as a free brain into any other harness or agent. So let's say, for example, you really like Hermes. Well, you could get a free brain like OmniRoot and start using Hermes as a separate profile with OmniRoot directly there as well. So it's a great system. And then additionally, this also plugs into your memory. So one of the biggest issues that you're going to find if you're using, for example, ChatGPT and the app over here, is that it doesn't sync, for example, with Cloud Code. Everything that you do with Cloud Code isn't synced up to ChatGPT. Why? Because they don't have the shared memory between them. However, if you have a system that plugs a memory system, a shared memory system into all of your AI agents and harnesses, like you can see right here, we're actually using Obsidian. And also, it looks super nice as well, so we can visualize it really easily. Then that will plug into Codex, into Hermes, into any other agent that we use, and it's automatically updated. So we can have a look inside our memory system here, and it's like, okay, 42 minutes ago, we did this, 54 minutes ago, we did this, three hours ago, we did this.