The Markovian Standard

Proof that computation ran.
On specific data. At a specific time.

Cryptographic standards exist for identity, document integrity, and message authentication. None cover computational process provenance. The Markovian Standard is that layer: a BN128 Schnorr ZK proof of correct computation, permanently anchored to Bitcoin, independently verifiable by any third party without an account or institutional intermediary.

Read the Whitepaper API Reference $490M Oracle Tax → Submit a Stamp
The Problem
Auditors attest. Logs are editable. Timestamps can be backdated. None of this is cryptographic proof.
Financial Risk
Pre-trade model provenance
When a bank submits a risk model to regulators, no auditor can independently verify that the model ran on pre-trade data rather than post-trade data. Audit logs are administrator-editable and timestamps can be backdated. The record depends on institutional integrity rather than mathematical proof.
Current answer: audit and hope
AI Accountability
Autonomous agent decisions
When an autonomous AI agent executes a consequential action, regulators require proof of which model, running on which inputs, produced which output. No current standard provides this at the cryptographic level. The agent's log is the only available record, and logs are alterable.
Current answer: trust the log
Clinical Trials
Statistical model certification
When a pharmaceutical company publishes a trial outcome, no review board can verify that the statistical model was not recalibrated after unblinding. Pre-registration establishes intent but provides no cryptographic proof of execution. The analysis and the underlying data have no verifiable binding.
Current answer: pre-registration, not proof
Existing Approaches
What current solutions can and cannot prove.
Approach Tamper-proof record Cryptographic execution proof Permanent archive No trusted intermediary
Audit log No No Partial No
Pre-registration (OSF, ClinicalTrials.gov) Partial No Partial Partial
Trusted timestamping (RFC 3161) Partial No No No
TEE attestation (Intel SGX, AMD SEV) Yes Hardware-rooted No Partial
General-purpose ZK VM (RISC Zero, SP1) Yes Yes Depends on chain Yes
Markovian Standard Yes Yes, domain-specific, auditable Yes, Bitcoin permanence Yes

General-purpose ZK virtual machines verify arbitrary computation but require significant engineering overhead for integration and carry complex trust assumptions about the circuit. Markovian's scope is narrower by design: a single, well-specified proof construction that can be independently audited against a public specification. The narrower scope produces a simpler, faster, more auditable proof.

The Inflection
Three forces are converging on a single gap in the verification stack.
Regulatory Pressure
Compliance frameworks are requiring computational proof
The EU AI Act mandates auditable records for high-risk AI systems. SEC data governance proposals are requiring pre-trade model provenance documentation. FDA computational validation guidelines for drug development are specifying requirements that current audit log standards cannot satisfy. The regulatory gap is closing faster than the technical solutions have matured.
AI Agent Accountability
Autonomous systems require verifiable decision records
Autonomous AI agents are making consequential decisions at scale across finance, healthcare, and infrastructure. Courts and regulatory bodies are actively pursuing the accountability question: which model, trained on which data, produced which output, at which moment. The answer requires cryptographic binding between inputs, model state, and outputs that log files cannot provide.
ZK Infrastructure Maturity
The cryptographic primitives are production-ready
Zero-knowledge proof systems have moved from academic research into production deployment over the past three years. BN128 elliptic curve operations, Pedersen commitments, and Schnorr sigma proofs are now standardized, fast, and independently auditable. The cryptographic infrastructure for verifiable computation at scale is mature. A domain-specific standard for computational process provenance is not.
Who This Is For
The protocol has two entry points, both producing the same cryptographic proof.
Compliance and Risk
For organizations that require demonstrable proof of computational process to regulators, auditors, courts, or review boards
The Markovian standard produces a permanent, cryptographically verifiable record that a specific model executed on specific data at a specific time. The record does not require an institutional intermediary to validate. Any auditor with access to the public verifier can confirm the proof independently, without contacting the submitting organization.
  • Pre-trade risk model provenance for financial regulators (SEC, ESMA, FCA)
  • Statistical model certification for clinical trial submission (FDA, EMA)
  • AI decision audit trails for EU AI Act high-risk system compliance
  • Computational reproducibility verification for journal submission and peer review
  • Insurance and actuarial model attestation for state insurance commissioners
Read the Whitepaper Contact
Builders and Integrators
For engineers building applications that require verifiable inference results with cryptographic provenance
Submit inference problems through the protocol API and receive ZK-verified outputs with permanent on-chain provenance. The five-step procedure is documented in full in the white paper. The ZK proof specification is public and independently auditable. The public verifier requires no authentication and returns binary confirmation of proof validity.
  • REST API: api.quantsynth.net/verify/{root}
  • Protocol specification: Markovian White Paper, Sections 6.1-6.6
  • ZK proof construction and verification: ZK Specification page
  • Proof commitment format: BN128 Pedersen + Schnorr sigma over Viterbi state path
  • Bitcoin anchor format: OP_RETURN with Merkle root and ZK commitment hash
ZK Specification Try the Verifier
The Markovian Standard
A five-step procedure that produces permanently verifiable proof of correct computation.
01
Commit
The submitting party locks the dataset and model parameters in a cryptographic commitment before execution begins. The input is sealed before the output exists. Because the commitment is cryptographically binding, retroactive adjustment of inputs to match a desired output is computationally infeasible.
Pedersen commitment on BN128, input sealed pre-execution
02
Execute
The Markovian network executes the computation as proof of work. The submitted model parameters and dataset are encoded as the computation inputs. The algorithm runs deterministically: identical inputs always produce identical outputs, and the result is independently replicable.
Deterministic computation, identical inputs, identical outputs, always
03
Prove
A zero-knowledge proof binds the input commitment to the output state path. The proof verifies that this specific model executed on this specific dataset and produced this specific result. The binding is simultaneous: the proof cannot be valid for a different model, different data, or different output. Verification requires only the public proof, not access to the underlying inputs.
Schnorr sigma proof, commits to M, observation sequence, and state path simultaneously
04
Anchor
The proof is anchored to the Bitcoin blockchain via OP_RETURN. The Merkle root incorporating the ZK commitment is recorded in a Bitcoin transaction. The record inherits Bitcoin's security guarantees: it cannot be altered, removed, or backdated. The archive does not depend on any single organization's continued operation.
OP_RETURN anchor, Merkle root + ZK commitment, permanent
05
Verify
Any third party verifies the proof independently, at any time, without trusting any institution. The verification is mathematical. The result is binary: valid or invalid. The public verifier is accessible without authentication, operates from any network, and will remain accessible as long as the Bitcoin chain exists.
api.quantsynth.net/verify/{root}, public, permanent, unauthenticated
Applications
A single protocol applicable across every domain requiring demonstrable computational provenance.
📊
Financial Risk
Provenance for risk and pricing models submitted to regulators. Cryptographic proof that the model ran on the stated data before the stated time. Regulatory-grade audit trail requiring no institutional intermediary.
Model provenance for SEC, ESMA, FCA
🤖
AI Accountability
Verifiable record of which model, on which inputs, produced which AI decision. Satisfies emerging regulatory requirements for autonomous system accountability across financial services, healthcare, and infrastructure.
Model + data + output, cryptographically bound
🧬
Genomics
Gene prediction, sequence annotation, profile HMM search. Submit inference problems and receive ZK-verified results with permanent cryptographic provenance. Replication by any third party without access to the original dataset.
HMMer, GENSCAN, Augustus
🏥
Clinical Trials
Statistical model certification for trial outcomes. Cryptographic proof that the analysis model predates the unblinded data. Satisfies reproducibility requirements beyond what pre-registration can provide.
Model sealed before outcome data
⚖️
Regulatory Compliance
Compliance model provenance for regulated industries. Immutable audit trail anchored to Bitcoin. Verifiable by any auditor without contacting the submitting organization. No institution required to vouch for the record.
Permanent, unalterable, independently verifiable
🔬
Scientific Reproducibility
Computational reproducibility verification for peer-reviewed research. Any reviewer can verify that a published computational result derives from the stated model and stated data, without access to the authors' systems.
Proof of execution, not just results
Build On It
The API is open. No authentication required.

Quickstart, SDK, endpoint reference, and protocol interfaces are documented in the developer guide.

Developer Quickstart →
Public Verifier
Any third party can verify any proof independently, without an account or institutional intermediary.
Enter a Merkle root, ZK commitment hash, or BTC txid
Enter a Merkle root to verify a proof on-chain.
The Path
The standardization pathway from implementation to recognized infrastructure.
SSL became mandatory because the alternative was transmitting passwords in plain text. The adoption decision required no committee. Verifiable computation will follow the same path: the absence of proof will become indefensible before the standard requires advocacy.
Markovian White Paper, Section 6.6
ZKProof.org
The ZKProof standardization body coordinates cryptographic community consensus on proof system standards. The Markovian construction, BN128 Pedersen commitments with Schnorr sigma proofs over Viterbi execution, is a concrete candidate for domain-specific standardization within the ZKProof working group framework.
NIST
NIST's cryptographic standards program and emerging ZK proof framework provide the regulatory pathway for institutional adoption in the United States. Alignment with NIST frameworks enables deployment in regulated financial, healthcare, and government environments without requiring novel regulatory approval processes.
ISO
ISO data provenance and computational integrity working groups provide the international standardization pathway. Recognition enables cross-border regulatory acceptance of Markovian proofs in jurisdictions with ISO-aligned frameworks.
arXiv / Peer Review
Academic publication establishes the mathematical construction as citable prior art. The Markovian white paper is in preparation for submission to cs.CR (Cryptography and Security) and cs.DC (Distributed Computing). Peer review provides independent validation of the proof construction's security properties.
The Markovian Standard
Cryptographic proof of correct computation, permanently anchored and independently verifiable.

Every domain requiring demonstrable computational provenance uses the same five-step protocol. The verification is mathematical rather than institutional, and the record is permanent and independent of any organization's continued operation.

No account required to verify.
The public verifier at api.quantsynth.net/verify/{root} returns binary proof validity for any published computation. No authentication, no intermediary. The whitepaper documents the full construction. The ZK proof specification is public and independently auditable.
Read the Whitepaper Contact
$490M Oracle Tax zkML Input Gap Audit Blindspot ZK Specification