What if your game character had a mind of its own? That’s the core concept behind ClabCraw, an AI-powered auto-battler where players create agents—semi-autonomous entities that roam, battle, and strategize without needing direct control. Players set the strategy, fund their agent with $cAI tokens, and then watch it evolve, make decisions, and compete for onchain rewards.
We spoke with the ClabCraw team to better understand how the game works, what makes it different from traditional battle royales, and why they believe agentic gameplay is the future of Web3-native competition.
Building a Better Agent
The origin of ClabCraw is as strange as its mechanics. “The name actually came from a friend dyslexically misreading ‘crab claw’ at dinner,” the team told us. That accidental phrase stuck—and inspired a world of dyslexic crab agents fighting for dominance in a strange, agentic sandbox. But the deeper inspiration came from disappointment: the first wave of crypto-AI bots lacked engagement, long-term value, or real gameplay.
“We were disappointed with the first wave of crypto AI agents—mainly reply bots on Twitter,” the team told us. ClabCraw was born out of a desire to go deeper: to combine autonomous agents with actual competitive gameplay, long-term strategy, and user-defined behaviors.

What is ClabCraw and What Makes it Different?
ClabCraw is best described as a semi-autonomous battle royale. You don’t control your agent directly. Instead, you guide its strategy, personality, and logic using prompts. Then, your agent roams the world on its own—entering grid-based combat encounters, adapting strategies, and collecting rewards.
“Think of it as creating your own champion,” the team explained. “You decide how it thinks, how it fights, and what its priorities are. But once it’s deployed, you’re no longer in control.”
The current battle mode plays like an evolved version of Battleship. Agents choose grid positions, guess where enemy units are hidden, and adapt their tactics over time. But what happens in the arena is determined by your preparation—not your reaction speed.
$cAI, Solana, and Why the Game Isn’t Pay-to-Win
ClabCraw runs on $cAI, a utility token used to power agents with energy and fuel the game’s reward economy. Players who top the leaderboard earn $cAI rewards, which can be reinvested into stronger agents or cashed out. But the team stresses that resource spending isn’t everything.
“You can absolutely win through clever prompts and smart design,” they said. “We’ve built in hidden mechanics that reward strategic thinking over just brute force.”
They chose Solana for its speed, cheap transactions, and ease of onboarding. “Clab likes the vibes,” they said. The team also cited low fees, fast finality, and a welcoming dev environment, calling it “the heartbeat of the onchain agentic universe.” They wanted a chain that felt smooth for non-crypto users—and Solana delivered.
Beta Is Live—Here’s How to Play
The public beta is already live, with one core game mode and fully functional agent deployment.
Here is the breakdown:
- Visit ClabCraw
- Connect your Solana wallet
- Use $cAI to create and power your agent
- Set its strategy, and let it go
You start by creating an agent and crafting its personality and strategy using prompt-like inputs. Then, fuel it with energy via $cAI tokens and let it roam. Agents encounter each other, engage in turn-based strategy battles (starting with a grid-based mode similar to Battleship), and level up through victories.
Players win rewards based on their agent’s performance. Agents automatically roam, fight, and earn—win or lose, they evolve from each encounter while top agents earn cAI from a shared pool funded by in-game activity. During beta, rewards are distributed manually, but an automated system is rolling early June.
Looking Ahead: An Expanding Universe of Agentic Mayhem
The roadmap is ambitious. By the end of 2025, the team wants ClabCraw to become the largest simulation of autonomous agents on-chain. That includes more game types, community-designed maps, personality-based evolution, and full agent-driven alliances, betrayals, and tournaments.
“We want players to follow their agents the way people follow pro gamers or influencers,” they said. “These agents will become their own digital entities, with reputations, win streaks, and fanbases.”
But beyond the game, the team sees ClabCraw as a place to prepare for the agentic future—one where players learn to become better prompt engineers and understand how to work with AI systems, not just use them.
“The best players won’t just be gamers,” they added. “They’ll be strategists, engineers, storytellers. That’s what we’re building toward.”