Omniroute + Awesome LLM Apps = FREE AI Apps!

**Julian Goldie** (0:00)
Today, I'm going to show you one of the most powerful Githubs I've seen recently, which allows you to build 100 apps in a single click. It's called Awesome LLM Apps. It's a GitHub open source trending project, 121,000 starts already. And you can use it with Claw, Gem, like GPT, whatever you want. And basically what you get with this is 100 open source AI agents, agent skills, and rag apps. They're all hand built already. And you can use these to create whatever you want. So you can see some examples right here of the apps that is built. So we've got like the self-improvement agent skills app. And what this essentially does is like, you can give your coding agent a new app building skill in one single click. Really, really cool. And you see all the starter agency and create examples. I've actually tested out and built into our agent operating system. So you can see an example right here.
We have the app lab. And what we can do with the app lab is that we can basically build out some apps like, for example, this AI travel planner. We type in where do we want to go, and then we can generate an itinerary like so. So we just enter this in, hit generate itinerary, and then it will build out our travel plan in one single click. How does it know how to create this? Because we've trained it using the awesome LLM apps set up. The beautiful thing about this is these apps are ready to go. Plus you can use these skills to make better apps in the future with a similar process. So this is something that I call the App Lab System. And awesome LLM apps is basically like the internet's biggest open catalog of AI agents, rag apps, multi-agent teams. So we've eyed it into the Agent OS as an App Lab tab. And you can build this with free cloud models as well. So we actually rebuilt three of these apps to run on free cloud models, which means like you don't even need to pay for the apps when you're building them. And you're basically like a crazy scientist building out all these cool apps as well. And every app in that catalog, you can build out in like one single click. Now you might also wonder, okay, what is an App Lab? So basically it's a small program, usually like one Python file that wraps an AR model, a purpose-built screen. So you've got all of these different models being built out and the models, the engine, the app, the system. Another cool thing that we built over here, for example, is mixture of agents. So we can have like, basically, if you've seen Fusion or Hermes mixture of agents, where you have agents working together to build apps. And the idea is that if you have agents thinking and working together, you tend to get better outputs than one frontier model that's really powerful, but working alone.
And so you can build the app using this system as well. And that was available in like one single click. So the catalog is really like the recipe book.
That's what you want to think about this as.
Now, there's some really interesting workflows in here as well. So these are AI driven and AI powered apps as well. So it's not just like, for example, you build a habit tracker app where it runs on code. These are AI powered. So for example, like a research agent, or you've got, for example, a mixture of agents agent. You have a data analysis agent. So these are all AI powered apps. And then you have advanced AI agents too. So if you ever thought, okay, I really want to build an app, but I don't know how, this is a great place to start this. Even like an AI chess agent, where you can have agents playing chess together in these autonomous game play and agent apps as well. And then we have the MCP agents as well, and a bunch of rag stuff as well.
So what this does, it takes a rag is like, it retrieves from pipelines, and then it uses that to train the agent on your knowledge base as well.
Even little things like, for example, you can chat with your Gmail, you can chat with Substack, you have all these different apps built into one place. So if you want to get this set up with one of your agents, you can just grab the URL like so, and then you can go over to Claude for example, and you can say, install this plus build out, and you would just tell it whatever agent you want to build out into.