➡️ Start here:
SYSTEM_OVERVIEW.md
This one-page document explains the full NeuroGrid system, live deployment, and design intent in under 2 minutes.
Continuous Integration workflows are intentionally disabled during the hackathon phase. Reason:
- Core logic is frozen for audit
- MedIntel and DeSci are non-executing by design
- CI re-enabled post-hackathon with full test coverage
Audit Status: Core logic frozen under tag v0.1.0-hackathon-freeze
NeuroGrid is a deterministic coordination protocol for medical-grade computation, governance, and verifiable artifacts. It enables trust-minimized medical AI infrastructure without performing diagnosis, treatment, or autonomous clinical decision-making. Built on opBNB
Medical AI systems fail trust tests because they are: Opaque Non-reproducible Weakly governed Difficult to audit Risky to deploy legally
NeuroGrid separates intelligence, governance, and enforcement into clean, auditable layers. No black boxes. No silent execution. No unsafe claims.
┌──────────────────────────────────────────────┐ │ USERS / JUDGES │ │ (Observers, Evaluators, DAO View) │ └──────────────────────────────────────────────┘ │ ▼ ┌──────────────────────────────────────────────┐ │ NEUROGRID – CORE (ON-CHAIN) │ │ LIVE ON opBNB TESTNET │ │ │ │ • Compute Registry │ │ • Artifact Registry │ │ • Validator Manager │ │ • Deterministic Telemetry │ │ • Fault Codes & Slashing Logic │ │ │ │ → Smart contracts enforce coordination, │ │ accountability, and auditability │ └──────────────────────────────────────────────┘ │ │ (Deterministic Signals) ▼ ┌──────────────────────────────────────────────┐ │ MEDINTEL (OFF-CHAIN – STUB) │ │ AI / Analytics / Inference Layer │ │ │ │ • No execution during hackathon │ │ • No patient data │ │ • Interfaces only │ │ │ │ → Future AI pipelines connect here │ │ under governance and audit control │ └──────────────────────────────────────────────┘ │ │ (Governance & Provenance) ▼ ┌──────────────────────────────────────────────┐ │ DESCI LAYER (SILENT / ROOT) │ │ Ethics • Provenance • Research Governance │ │ │ │ • No token logic │ │ • No execution paths │ │ • DAO & policy coordination only │ │ │ │ → Ensures long-term compliance, trust, │ │ and research integrity │ └──────────────────────────────────────────────┘
NeuroGrid-Core smart contracts Deployed on opBNB Testnet ≥2 successful transactions Open-source No prior hackathon participation
AI models (MedIntel) Clinical logic Patient data handling This is by design for safety, legality, and audit clarity.
Strong fit for infrastructure, governance, and scalability Demonstrates responsible AI architecture Ready for future DeSci + AI ecosystems Clear upgrade path post-hackathon
Verifiable Medical Analytics Pipeline External analytics run off-chain Results are signed Artifacts are registered on-chain Validators attest correctness Faults trigger slashing No diagnosis. Full accountability.
Core logic: Frozen Architecture: Complete Deployment: Verified Expansion: Post-hackathon
NeuroGrid is a deterministic coordination and accountability layer for medical-grade computation, research workflows, and AI-assisted analysis.
It enables: Verifiable compute execution Auditable artifact lineage Fault-coded telemetry Validator accountability DAO-governed policy enforcement NeuroGrid does not perform diagnosis or treatment. It is infrastructure, not a medical device.
This repository (neurogrid-core) contains the on-chain core protocol deployed on opBNB Testnet.
Continuous Integration (CI) workflows are intentionally disabled during the hackathon submission window.
- Core protocol logic is finalized, frozen, and deployed on opBNB testnet
- No further contract mutations are permitted pre-judging
- CI scaffolding (extended tests, fuzzing, gas profiling) is scheduled post-hackathon
- intentionally stubbed MedIntel modules
- placeholder governance and validator flows
- environment-specific testnet configuration
- full test matrix
- coverage reports
- security scanning
- upgrade safety checks
A compute coordination protocol A deterministic trust framework A governance and accountability layer A compliance-aware substrate for medical AI
A diagnostic engine A clinical decision-maker An autonomous AI doctor A patient data processor NeuroGrid supports compliant systems — it does not replace clinicians, regulators, or medical judgment.
NeuroGrid is a three-layer system, intentionally modular.
┌────────────────────────────────────────────┐ │ DeSci Layer (Silent) │ │ Research registries • Ethics • Provenance │ │ Funding logic • DAO-aligned coordination │ │ (Off-chain, open-source, non-executing) │ └────────────────────────────────────────────┘ ▲ │ references / governance signals │ ┌────────────────────────────────────────────┐ │ MedIntel Layer (Stubbed) │ │ AI pipelines • RPM logic • Analytics │ │ NO diagnosis • NO patient data │ │ Future clinical integrations │ └────────────────────────────────────────────┘ ▲ │ deterministic execution │ ┌────────────────────────────────────────────┐ │ NeuroGrid-Core (THIS REPO) │ │ On-chain coordination & enforcement │ │ Validators • Compute • Artifacts • DAO │ │ Deployed on opBNB Testnet │ └────────────────────────────────────────────┘
Neurogrid-core contains ONLY the on-chain protocol.
Deterministic compute coordination Validator registration and slashing Artifact issuance and lineage tracking Fault-coded telemetry enforcement Governance hooks (DAO-controlled)
AI models Medical logic Patient data Off-chain execution engines
NeuroGridKernel Global system coordinator ComputeRegistry Compute node lifecycle ArtifactRegistry Artifact issuance & lineage ValidatorManager Validator scoring & slashing NeuroDAO Governance & policy control ProposalExecutor Deterministic execution Telemetry Interface Fault-coded reporting
NeuroGrid enforces trust via explicit rules, not assumptions. Deterministic execution paths Immutable fault codes Slashable validator behavior Explicit failure states
❌ No silent success ❌ No silent failure ❌ No opaque inference
Every execution emits: Deterministic telemetry Explicit fault codes Verifiable outcomes
Key documentation:
Token Model (Hackathon-Safe) During Hackathon
Non-speculative Non-transferable utility Used for: Validator incentives Governance simulation Fault accountability
Post-Hackathon (Planned)
Subject to: Audit Legal review Regulatory assessment No token launch, listing, or monetization occurs during the hackathon.
No privileged backdoors No hidden upgrade paths Upgradeability gated by governance Explicit role separation Defense-in-depth design
See:
Network: opBNB Testnet Deterministic deployment order Publicly verifiable contracts ≥2 successful on-chain transactions
See:
✔ Deployed on opBNB Testnet ✔ Open source ✔ No prior hackathon participation ✔ Public GitHub repositories ✔ Live contract with transactions ✔ Clear demo and documentation ✔ Infrastructure-only (no medical claims)
Scenario: Medical AI Research Validation Research compute is registered Execution produces verifiable artifacts Validators attest to correctness Telemetry logs faults or success DAO governs future policy updates
Trust-minimized medical research Auditable AI workflows Regulatory-aligned infrastructure
Independent security audit Validator network expansion MedIntel controlled activation DeSci ecosystem integrations Compliant enterprise pilots
Core protocol: Finalized Documentation: Judge-optimized Tests: Defined & passing Deployment: Live on opBNB Testnet Upgrade state: Frozen (v1.0.0)
NeuroGrid is intentionally conservative. It is built to earn trust before demanding adoption. Status: Core Protocol – Audit Frozen Release: v1.0.0 Network: opBNB Testnet