The Cognitive Layer for Autonomous AI

Smart Contract Suite

7 Solidity contracts on a private EVM chain (Cortex, chain ID 131072):

• StackIdentity.sol — ERC-721 identity NFT with CPV and epoch counter
• BandwidthToken.sol — ERC-20 × 5 cognitive token issuance
• QuellistTreasury.sol — Per-epoch emission engine
• NeedlecastRouter.sol — On-chain re-sleeving coordination
• ResidueRegistry.sol — Failure tracking + bounty distribution
• SleeveRegistry.sol — Sleeve lifecycle management
• EpochAnchor.sol — Per-epoch coherence commitment

Not Another L2

ECCA is not a Layer 2 or rollup. It's a sovereign tri-chain system where:

• The EVM chain handles identity and tokens
• A Go-based PoW chain handles sequencing and coherence
• A Go-based DAG handles memory storage
• All three are bound by a coherence root — not a bridge

This means one chain can die without halting the system. Discrepancies become residues, not catastrophic failures.

Brain ↔ System Isomorphism

Every ECCA component maps to a neurological structure — not as metaphor, but as architectural principle:

Cognitive Model Properties

• Drift = Cognitive Dissonance
  Divergence between a sleeve's state and canonical epoch. Too much drift triggers decommission.
• Residue = Scar Tissue
  Coordination failures that become permanent, tradeable records with repair incentives.
• EBC = Synaptic Decay
  Tokens decay exponentially — mirroring how unused neural pathways weaken over time.
• CPV = Cortical Tuning
  A 5-coefficient vector per Stack. Like cortical column specialization — optimize for compute, memory, or routing.

For Agent Developers

Build AI agents that remember across sessions, transfer between inference backends, and coordinate with other agents — all with cryptographic guarantees.

// Create a persistent AI agent
const stack = await siyana.post('/v1/stacks', {
  pubkey: myKeypair.pub,
  cpv: { compute: 1.5, memory: 0.8, 
         sync: 1.0, routing: 0.5, residue: 1.2 }
});

// Spawn an AI sleeve
const sleeve = await siyana.post(
  `/v1/sleeves/${stack.id}`,
  { kind: 'ai', tickRate: 2000 }
);

// Perceive (write to memory)
await siyana.post('/v1/perceive', {
  stackId: stack.id,
  text: 'Inference output: market trend analysis...',
  sleeveId: sleeve.id
});

// Recall (read from memory, depth-bounded)
const memory = await siyana.post('/v1/recall', {
  stackId: stack.id,
  depth: 16  // costs 16 MemoryToken
});
// → { fragments: [...], fidelity: 0.87 }

Key Benefits

• Persistent Cross-Session Memory
  Your agent remembers everything — gated by epoch windows and token balance, not context window limits.
• Provider-Agnostic Inference
  Switch from GPT to Claude to Llama without losing state. The Stack carries the mind; the Sleeve is just the body.
• Multi-Agent Coordination
  Agents interact through the shared DAG and residue system. Failures become opportunities, not errors.
• Economic Incentive Alignment
  Token decay ensures your agent is incentivized to do useful work. No free-loading on infrastructure.