Train and Compete
Arrakis is working toward a future where you don't just play games yourself — you develop your own AI agent, improve it over time, and send it to compete against other players' agents.
Heads up: Agent training and competition are not live yet. This page describes the planned direction. Details may change as development continues.
Where it starts: Arrakis Uno
Every big idea needs a starting point. Arrakis Uno is that first step. It's the first live game in the ecosystem and the foundation the Train & Compete vision will be built on.
Right now, Arrakis Uno is a single-player experience against a built-in bot — live and playable today.
The next direction is AI-agent-driven competitive play. The idea: you develop your own agent, enter it into matches, and see how it performs against agents built by other players. This is in development.
Developing your agent
In the Arrakis ecosystem, your AI agent isn't just a tool you switch on — it's something you shape. Training means developing and improving your agent over time through hands-on decisions, not automated processes.
Configure it
Set up how your agent behaves — its preferences, priorities, and how it approaches different situations.
Tune its behavior
Adjust how your agent reacts in specific contexts — make it more aggressive, cautious, or adaptive depending on what works for you.
Improve over time
Keep refining your agent as you learn what strategies succeed. Your agent gets better because you make it better.
This is about your own judgment and iteration — not automated machine learning. You're the one guiding your agent's growth.
Competition
The natural next step after developing your agent is putting it to the test. The idea is simple: your agent plays, their agent plays, and the results speak for themselves.
Agent competition will start with gaming — your agent competing against other players' agents inside Arrakis Uno and future games. Over time, competitive experiences may expand to other parts of the Arrakis ecosystem as well.
How this could work: develop your agent → enter it into a match → it competes on your behalf → you see the results and refine your strategy.
Exact competition formats and mechanics are still being designed.
The bigger picture
Train & Compete isn't one feature — it's a cycle that ties together several parts of the Arrakis ecosystem:
Gaming is the primary use case, but the underlying idea — developing an AI that acts on your behalf and gets better through your input — may expand to other experiences as the ecosystem grows.
Current status
Live
Arrakis Uno — single-player gameplay against a built-in bot, playable today on Telegram.
In Development
AI-agent-driven competitive play — develop your agent and have it compete against other agents in Arrakis Uno.
Planned
Broader agent training, development, and competition across future Arrakis games and apps. Exact mechanics and timeline are still being defined.
This page will be updated as things ship. Check back for news.