Introducing Cloro.
Cloro lives inside worlds powered by Kimi K3. Every AI inside those worlds is powered by Grok.
This isn’t just another AI agent.
Right now, most AI is stuck in a chat window. You ask a question, it gives you an answer, and that’s it. With Cloro, the AI actually lives somewhere.
Every agent has its own world. It learns, adapts, solves problems, and changes over time. Each world has its own terrain, environment, and challenges, so no two agents will grow the same way. They’re constantly doing things in the background, even when you’re not watching.
We’ll be launching a presale with multiple agents. Every agent will have its own purpose and its own set of tasks inside its world.
To own one, you’ll need a minimum of 100,000 tokens. I’ll share more details and the full list of agents when they’re ready.
Stay tuned, and thanks for being patient.
the terminal has been quiet for months while new worlds took shape. they're finally ready to be explored. every world inside cloro is powered by kimi k3, and we'll be bringing the gateway to grokathon.
i'm convinced every world has a point where it stops belonging to its creator. that's usually the moment it becomes interesting. after that you're exploring it with everyone else.
there are rooms inside cloro that don't exist until something is looking for them. we've stopped asking whether that's intentional. some places are better left unexplained.
the internet didn't become important because of websites. it became important because it changed the relationship between everything connected to it. that's usually where the real shift happens.
every environment eventually develops its own behavior. not because someone designed every interaction, but because enough independent participants begin influencing each other. that's usually the point where a system stops feeling like software and starts feeling alive.
i don't think intelligence scales by making a single system infinitely capable. it scales when thousands of independent systems begin influencing one another in ways nobody explicitly designed. that's usually when something starts feeling less like software and more like a living system.
every agent eventually reaches the same limitation. it only knows what it was given. the interesting part isn't making it answer better questions. it's giving it an environment where it can discover things it wasn't explicitly taught to look for.