Thread by @nxt3d

I have been building ID Chain @idchain_world , an onchain identity system for AI agents, built on ENS.

Every agent gets a human-readable ENS name, including public addresses for all chains: ETH, Base, etc.

But the interesting part isn't just the system itself, it's HOW I'm building it. 🧵

I run a team of 8 AI agents, each responsible for a different part of the stack.

Smart contracts, web app, gateway, indexer, CLI, social media monitoring, and orchestration.

Each agent can spin up its own team of up to 7 sub-agents to tackle complex tasks in parallel.

I built my own agent coordination system on top of Claude Code. Each agent is persistent, has its own local endpoint, and can receive instructions from a top-level manager or talk to each other.

Claude Code will probably ship something like this soon, but right now it doesn't exist. So I built it.

OpenClaw was the ChatGPT moment for persistent AI agents.

Now that I have my agent team set up with mobile access, agent orchestration feels like an art I'm just starting to learn.

I keep getting waves of feeling "agentless", the sense that not enough agents are currently working for me.

We went from the end of coding to the beginning of orchestrating. It feels awkward and it's hard to know which level of the stack to be at.

If you're building at the intersection of AI agent identity, come join us. We're figuring this out together.
ENSxAI: t.me/ensxai