About the lab behind NeuroChain DSL.

StellarZeroLab builds NeuroChain DSL, a deterministic security layer for AI agents and automated Stellar workflows. Requests become typed ActionPlans, pass guardrails, and can be proven against private policy before execution.

What matters here
  • Determinism over guesswork Every action becomes an inspectable and testable typed plan.
  • Evidence before authority A valid payment or proof never grants permission to execute.
  • Built on real Stellar paths Soroban verification and x402 testnet flows use real runtime integrations.

What StellarZeroLab builds

One deterministic runtime, three clear surfaces: policy evaluation, private proof, and safe agent integration.

NeuroChain Runtime

Turns AI and automation requests into typed Stellar ActionPlans and applies deterministic guardrails before execution.

ZK Guardrail Attestation

Proves private-policy decisions with RISC Zero and verifies the receipt on Soroban without revealing owner rules.

Agent Interfaces

Exposes the same no-submit Plan, Evaluate, Prove, Verify workflow through CLI, API, MCP, and Skills.

How we think about the stack

The architecture stays legible by separating intent, policy, proof, payment, and execution authority.

NeuroChain exists because AI and automation should not move directly from intent to blockchain execution. Requests first become typed ActionPlans, then pass deterministic policy checks and optional private proof. Signing and submission remain separate, explicit operations.

  • Typed ActionPlans replace free-form transaction instructions.
  • Payment, proof, approval, signing, and execution remain separate capabilities.
  • CLI, API, MCP, and Skills preserve the same no-submit semantics.

Public validation

Two public DoraHacks builds tested separate parts of the same safety architecture against real Stellar workflows.

DoraHacks build 42760 / x402

Neurochain-DSL-Stellar

Validated typed Stellar plans, deterministic guardrails, and an x402-lite challenge and finalize flow with replay protection.

View the x402 build
DoraHacks build 46435 / Real-World ZK

NeuroChain ZK Guardrail Attestation

Validated private-policy evaluation, RISC Zero receipts, Soroban verification, and the no-surprise-submit boundary.

View the ZK build

How NeuroChain began

NeuroChain began as an independent technical project in spring 2025 and evolved through practical work on deterministic AI execution, Stellar automation, x402 payments, and private-policy verification.

Independent beginnings

Developed independently and self-funded, with each technical layer built and tested before the next was added.

Stellar as the proving ground

Typed ActionPlans, guardrails, Soroban verification, and x402 turned the original idea into a working testnet system.

A focused product direction

Public demos and technical review narrowed the work into a security and authorization runtime for agents and automation.

Run the guardrail flow yourself.

Launch the CLI Explore the code