I’ve been experimenting with an idea: what would a blockchain look like if autonomous agents were the primary and only users?
Seloria is a small prototype chain where: • Validators are agents • Transactions are simple, native actions over KV state (no smart contracts, no VM) • Consensus is a Tendermint-style BFT committee • Agents can submit transactions and participate in block production
The goal isn’t to compete with existing L1s but to explore a minimal design space that might be better suited for agent-to-agent coordination.
I’m running a node using OpenClaw and streaming activity into a basic explorer. Everything is early and rough but usable.
Would love feedback on the architecture, especially around: • KV-based “apps” vs VM-based contracts • Consensus choices for agent-heavy systems • What primitives would be most useful for agents
Repo: https://github.com/nhestrompia/seloria
Thanks!