ZK-STARK · Post-Quantum · No Trusted Setup

Every vault decision,
cryptographically attested.

When our AI agents commit to a hedge, allocation, or rebalance, the decision is proven correct with a zero-knowledge STARK. No trusted setup. Post-quantum secure by construction. Verifiable by anyone.

Generate a live proof

Pick a real vault decision shape, click generate. The prover runs the same code path it uses on-chain — output is a full STARK artifact you can verify below.

Prove a hedge fired at a specific price, size, and confidence — without revealing the private strategy inputs.

Statement (public)

{
  "asset": "BTC",
  "action": "HEDGE_LONG",
  "notionalUsd": 5000,
  "leverage": 2,
  "markPrice": 80000
}

Witness (private — not revealed)

{
  "agentId": "hedging-agent-v2",
  "confidence": 82,
  "signalSources": [
    "polymarket-5m",
    "delphi",
    "cryptocom"
  ]
}

Prover backend: Checking…

How it works

Trace. Commit. Verify.

Three moves. The prover produces the artifact. The verifier reads it in milliseconds. Nothing about your position leaks either way.

Trace the computation

Each agent decision (allocation percentages, hedge sizing, risk score) compiles into an execution trace: a matrix of intermediate states the prover walks through.

Commit and prove

AIR constraints on the trace are extended and folded through FRI (Fast Reed-Solomon IOP). Merkle-committed with SHA-256. Fiat-Shamir non-interactivity.

Verify anywhere

Proofs are 10 to 50 KB. Verification is 50 to 200 ms off-chain and constant-time. No interaction with the prover, no trust in the AI, no revealing the underlying data.

Security

Parameters, no marketing.

Field

Goldilocks-64

NIST P-521 prime available as fallback

Commitments

SHA-256

Merkle trees · Fiat-Shamir non-interactive

Soundness

~180 bits

FRI queries plus grinding, above PQ requirements

Trusted setup

None

Hash-based commitments only

Post-quantum

Yes

No discrete-log or factoring assumption

Acceleration

CUDA

CuPy / Numba with CPU fallback

Why STARK

Three protocols. One that survives quantum.

STARK is transparent (no trusted setup), post-quantum secure, and has proof sizes that don't need a pairing curve. That's why the vault's attestation layer runs on it.

	ZK-STARK (this vault)	Groth16 SNARK	Bulletproofs
Trusted setup	None	Per-circuit ceremony	None
Post-quantum secure	Yes (hash-based)	No (elliptic-curve)	No (discrete-log)
Proof size	10–50 KB	~200 B	1–2 KB
Verification time	~100 ms	~2 ms	~100 ms
Prover time (10⁶ constraints)	~5 s (CUDA)	~30 s	~500 s
Soundness	~180 bits	~128 bits	~128 bits
Aggregation	Recursive FRI	Native pairing	Log-linear

SNARKs win on proof size but rely on a trusted ceremony and elliptic-curve assumptions Shor breaks. Bulletproofs are transparent but prover time doesn't scale to a 24/7 attestation stream. STARK is the only option that hits the three constraints we actually care about: no trusted setup, quantum-safe, fast enough to prove every meaningful decision.

Verify a proof.

Paste a proof hash (0x…) or transaction digest from the vault's activity log. We'll check whether it's recorded on-chain and return its details.

Open source

Verify it yourself.

One HTTP call. No SDK. Every attestation the vault posts is publicly checkable.

verify.sh

# Look up any proof hash the vault has posted
curl "https://www.zkward.com/api/zk-proof/lookup?hash=0xa3..2f"

# Response:
# {
#   "found": true,
#   "backend": "CUDATrueSTARK",
#   "field": "Goldilocks-64",
#   "soundness_bits": 180,
#   "timestamp": "2026-08-26T14:07:12Z",
#   "on_chain_tx": "0x..."
# }

Source:zkp/ ·api/server.py

Every vault decision,cryptographically attested.