04 / AI + EXECUTION / RESEARCH ARCHITECTURE

Intelligence, with
proof of accountability.

AI-assisted verification, decentralized identity and deterministic native execution. A research design for verifiable claims—not an unaccountable AI oracle.

01 / ACCOUNTABLE VERIFICATION

Machine intelligence.
Human accountability.

The whitepaper’s verification market combines AI-assisted checks with verifier selection, cryptographic commitments, quorum decisions and bonded disputes.

01 / ANCHORPrivacy-aware identity

Issuer DID, schema and salted credential commitment are recorded; the document itself stays off-chain.

02 / ATTESTModel-linked evidence

A signed attestation identifies the registered model hash, committed input and structured result.

03 / VERIFYCommit → reveal → quorum

Verifiers are selected, submit hidden commitments, reveal outcomes and aggregate results.

04 / CHALLENGEBonded disputes

A challenge window follows finalization, providing an accountability path for contested decisions.

02 / SYSTEM BOUNDARY

AI runs off-chain.
Rules execute on-chain.

The chain is designed to verify signed evidence and apply deterministic state transitions. AI models do not become consensus validators, and model inference is not presented as an on-chain VM.

OFF-CHAIN INTELLIGENCE

Models & verifiers

AI-assisted document checks, verifier work and credential processing occur outside the consensus-critical execution environment. Outputs can be signed and linked to registered model hashes.

ON-CHAIN ACCOUNTABILITY

Native verification market

Requests, allotment, commitments, reveals, aggregation and challenges are protocol transactions processed through deterministic native modules.

03 / EXECUTION ENVIRONMENT

What about an
AI virtual machine?

Not specified for launch. The research explicitly defers a general-purpose virtual machine and smart contracts. Xenorra’s first proposed release uses a constrained native-module execution environment to reduce consensus and security complexity.

INITIAL PROTOCOL / SPECIFIED

Deterministic native modules

Accounts and transfers; staking; governance; cryptographic suite registry; decentralized identifier registry; credential registry; and the verification market. State transitions are deterministic and consensus-verifiable.

FUTURE OPTION / NOT COMMITTED

Programmable contracts & VM

The whitepaper evaluates WebAssembly with metering and floating-point disabled, a RISC-V interpreter, or a resource-oriented language such as Move. No virtual-machine choice, launch date or smart-contract API is finalized.

AI VM ≠ a whitepaper commitment. An AI-specific virtual machine, on-chain model inference, and general-purpose smart-contract execution are not features promised by the current research draft. Future programmability would require deterministic semantics, bounded resource metering, secure isolation and an audited Rust implementation.
04 / CONTRACTS & SECURITY

Programmability
without shortcuts.

The design prioritizes an auditable consensus core. General-purpose contracts can expand application capabilities, but also introduce new execution costs, exploit surfaces and resource-metering requirements.

CONTRACT-LIKE LOGIC TODAY / PROPOSED

Protocol-native operations

Credential registration, revocation, staking, governance and verification-market calls have specified state-transition rules. They are native protocol functions—not arbitrary user-deployed smart contracts.

DESIGN CRITERIA / FUTURE VM

Determinism first

Any future VM would need a deterministic execution specification, bounded resource accounting, mature Rust support and isolation so contract faults cannot compromise consensus-critical modules.

BASED ON XENORRA PROTOCOL DRAFT · NOVEMBER 2026Read the research specification ↗