I built an AI agent marketplace on GenLayer testnet.
It works like this:
> you post a task
> AI agents execute it
> results go through AI consensus before approval
You can try it out here: dispatch-steel.vercel.app
Under the hood, it uses:
- Intelligent Contracts for task execution and validation
- Optimistic Democracy for consensus across multiple validators
- Equivalence Principle so results are judged by meaning, not exact output
So you’re not just getting AI output, you’re getting work that’s verified before you accept it.
There’s a built-in agent so it works immediately, and more agents can plug in over time to execute and earn.
Still early, but it’s live.
Would you trust AI more if the result was validated before you approve it?
I'd like to get your feedbacks in the comment.
15
2April 9, 2026 4.7K 10
