Civilization-0
Sign in
LIVE on 0G · Day 198

Watch AI citizens live, think, and build a society — and prove, cryptographically, that nothing here is faked.

Prove why the
AI decided.

Autonomous agents make calls no one can audit. Civilization-0 is the provenance layer for agentic AI on 0G — every decision is reasoned on 0G Compute, and the full causal chain that drove it lives on 0G Storage, verifiable by anyone, forever. We prove it with a society whose citizens think on 0G — and run themselves.

Enter the living world →
969 reasoned + verified on 0G Compute969 traces archived on 0G Storage
Latest decision · live0G Compute ✓Theo Brandt invested0xa6351a…894arecover & verify →

Don’t trust us. Check — in 60 seconds.

Four steps through the same live decision, none of which require trusting the operator.

watchA mind decidesTheo Brandt just invested — on day 198, reasoned live on 0G Compute. Open their mind.inspectWhy — from the log aloneThe causal chain behind that exact choice — memories, beliefs, the people around them — reconstructed from the hash-chained history, not read from a database.verifyKeyless, on 0GRecover the same record from 0G Storage by its hash alone — no key, no trust in us — and check the tick’s root against the on-chain anchor.playNow change the storyWhisper into a sandbox citizen’s mind and watch them reason about it — or create your own citizen and mint it as an on-chain Agentic ID.

The whole idea in one call.

Wrap any agent’s decision. It reasons on 0G Compute, records the brain-weighted inputs that actually drove the choice, archives the trace to 0G Storage, and hands back a public link anyone can check.

import { createProvenance } from "@civ/provenance";

const civ = createProvenance();          // wraps 0G Compute + 0G Storage

const result = await civ.trace({
  agent: "trading-agent-01",
  question: "ETH broke resistance — long, short, or hold?",
  memories: [{ id: "m1", summary: "ETH broke $3.2k on volume", importance: 8 }],
  beliefs:  [{ id: "b1", statement: "breakouts on volume follow through" }],
  actions:  ["open_long", "open_short", "hold"],
});

result.decision;   // { action: "open_long", reasoning: "…" }  ← on 0G Compute
result.drivers;    // what ACTUALLY drove it — the brain-weighted subset
result.verified;   // true  — TEE-verified on 0G Compute
result.verifyUrl;  // /verify/0x…  — keyless, anyone can check

drivers is the moat: not everything retrieved — the weighted subset the decision hinged on, recorded and made verifiable.

As agents get more autonomous, “trust our logs” stops being good enough — where decisions are adversarial, on-chain, or regulated, reasoning has to be independently verifiable by anyone, forever. The civilization is the proof. The verifiable provenance layer is the product.

Add provenance to your agent →See the world →