The full stack · built in the sprint

Nine interlocking pieces.
Three open. Six are the moat.

DCS is not one product. It is three open standards anyone can adopt — and six features that compose into one provenance-and-payments substrate for the AI agent economy. Build one, you have a feature. Build all nine with cryptographic linkage, you have a category.

3
Open standards — R+2 / R+3 / R+4. MIT-licensed. Open for independent review by standards bodies.
6
Production features — Trust, A2A, Memory, Economy, Federation, Self-Healing.
~3,700
Lines of backend route code across the 6 features.
Layer 1 · the open standards
Public good · MIT

Three standards — anyone can adopt them

We give the standards away. They are open specifications, MIT-licensed, with reference implementations. Adoption by one major lab or government brings the whole stack into worldwide use.

Layer 2 · the moat
6 features

Six features — the part nobody can clone fast

The standards are open. The reference infrastructure is the moat. Identity + memory + negotiation + economy + federation + self-healing, all cryptographically linked, is a longer substrate project for anyone starting today; timeline estimates are illustrative, not guarantees. We shipped the linked reference substrate quickly on low infrastructure cost. Production hardening, multi-peer federation, and independent audit remain open gates.

#5 · Trust

On-Chain Reputation + Trust SKU

DCSWorkerSBT soul-bound identity on Base mainnet + Trust Score + receipt chain, packaged as a Trust-as-a-Service SKU.

LIVE · ON BASE MAINNET
View on Basescan →
#2 · Memory

Sovereign AI Memory Graph

pgvector-backed memory, 1536-dim embeddings, ed25519-signed receipts, multi-tenant, auditable append-only audit log.

LIVE
View memory →
#38 · Federation

Global Agent Federation v1

Cross-tenant, cross-border agent routing with PKI for organisational identity. GDPR / DPDP / PDPL compliance engine built in.

LIVE · HANDSHAKE VERIFIED · CURRENT PEER COUNT = 1 · MULTI-PEER PRODUCTION FEDERATION NOT YET CLAIMED
View federation →
#6 · Negotiation

Agent-to-Agent Negotiation

Multi-round LLM bargaining, constraint schemas, R+2 receipts on every action. On-chain escrow + USDC settlement running on Base Sepolia testnet.

NEGOTIATION LIVEON-CHAIN ESCROW/SETTLEMENT · TESTNETBASE MAINNET MONEY MOVEMENT FLAG-GATED UNTIL INDEPENDENT SECURITY REVIEW
View A2A →
#70 · Economy

Multi-Agent Economy Platform

Agent wallets, service registry, reputation-weighted discovery, transaction ledger. On-chain USDC settlement running on Base Sepolia testnet.

LEDGER LIVESETTLEMENT · TESTNETMAINNET MONEY MOVEMENT FLAG-GATED UNTIL INDEPENDENT SECURITY REVIEW
View treasury →
#13 · Self-Healing

Self-Healing Infrastructure

Anomaly detection, auto-remediation, multi-region failover, cost-report cron. The reliability layer that keeps the stack up.

LIVE · HEALTH PROBES ACTIVE · AUTO-REMEDIATION DRY-RUN BY DEFAULT
View security →
why this is defensible

The moat

Copying one interface is a feature. Reproducing the linked identity–memory–negotiation–settlement–audit chain with working cryptography is a longer substrate project. Timeline estimates are illustrative, not guarantees.

A competitor can copy the R+2 receipt format in a weekend — it's open, that's the point. What they cannot copy quickly is the cryptographic linkage between all nine pieces: identity that resolves to the signing key, memory that emits signed receipts, negotiation that reads reputation, an economy that settles against identity, federation that carries compliance across borders, audit export that aggregates everything, and zero-knowledge proofs that verify it all without disclosure.

Provenance without identity is unaudited. Identity without memory has no history. Memory without provenance is unverifiable. An economy without provenance is uninsurable. Each piece needs the others to be meaningful — which means a competitor has to build all of them, correctly, with the cryptography linking up. That is a longer substrate project; timeline estimates are illustrative, not guarantees. We shipped the linked reference substrate quickly on low infrastructure cost. Production hardening, multi-peer federation, and independent audit remain open gates.

That gap is the moat. The standards stay open so the category grows; the linked infrastructure stays ours so the category has a home.