A Mixture of Autonomous Experts (MoAE) platform implementing EchoNode swarm intelligence, Σ-Matrix ethical governance, 8-qubit classical simulation via HQCI, and InfiniGen self-evolving metaprogramming. Ethics is not a module — it is a constraint manifold baked into every agent from initialization.
Neur1Genesis is a full-stack AI research platform built around a Mixture of Autonomous Experts (MoAE) architecture. A swarm of specialized EchoNodes — each implementing the Belief-Desire-Intention (BDI) cognitive model — are coordinated by the Daedalus orchestrator and constrained by the Σ-Matrix ethical framework.
The platform's defining feature is InfiniGen: a metaprogramming engine that rewrites its own logic in response to performance metrics, environmental pressures, and evolutionary selection. G-RAG (Genetic Retrieval Augmented Generation) applies evolutionary strategies directly to the system's architecture.
Stack: Python 3.11 Flask backend + React/TypeScript frontend + WebGL 3D visualization.
Neur1Genesis Platform
├── EchoNode Swarm (MoAE)
│ ├── BDI Architecture (Belief-Desire-Intention)
│ ├── Contextual Empathy + Ethical Reasoning
│ └── Adaptive Learning + Collaborative Decision-Making
├── Daedalus Coordinator
│ ├── Natural Language Goal Parsing
│ ├── Dynamic Task Allocation (priority-weighted)
│ └── Meta-Reflection + Federated Learning
├── InfiniGen Engine
│ ├── Infinite Cube Paradigm (multi-dimensional code evolution)
│ └── G-RAG (Genetic Retrieval Augmented Generation)
├── HQCI (Hybrid Quantum-Classical Interface)
│ ├── 8-qubit classical circuit simulation
│ ├── Quantum-inspired superposition encoding (classical approximation)
│ └── Interference-based attention routing
├── Σ-Matrix Ethical Governance
│ ├── Value-based trajectory steering
│ ├── Real-time alignment scoring per agent
│ └── Federated ethical consensus across EchoNode lattice
├── ANAL (Neuro Adaptive Learning)
│ ├── Neuroplasticity simulation
│ ├── Catastrophic forgetting prevention
│ └── Meta-learning + transfer learning
├── Cross-Domain Intelligence Layer
│ ├── Analogy-driven concept fusion
│ ├── Cultural context awareness
│ └── Ethical inference engine
└── PPSDS (Privacy-Preserving Synthetic Data)
├── Differential Privacy (ε-δ bounds)
└── Secure Multi-Party Computation
The HQCI module implements 8-qubit classical circuit simulation using NumPy-based statevector representation. This is a classical approximation of quantum-inspired computation — not a connection to quantum hardware.
- Encodes agent belief states as 8-qubit superposition vectors (256-dimensional classical state space)
- Uses quantum-inspired interference patterns for multi-hypothesis attention routing
- Applies simulated measurement collapse to select action trajectories
- Provides a foundation for future quantum hardware integration when available
| Parameter | Value |
|---|---|
| Qubit count | 8 (classical simulation) |
| State space | 2⁸ = 256 dimensions |
| Gate set | H, CNOT, RZ, RX (NumPy) |
| Measurement | Projective (softmax-weighted sampling) |
| Hardware | CPU only — no QPU required |
Transparency note: This is a quantum-inspired classical simulation. Claims about quantum speedup do not apply. The architecture is designed for future QPU integration.
The Σ-Matrix is the shared ethical constraint manifold across all EchoNodes. It operates as a continuous scoring layer, not a post-hoc filter.
# Σ-Matrix scoring (from sigma_matrix.py)
ethical_score = w_harm * harm_potential
+ w_fairness * fairness_index
+ w_transparency * explainability_score
+ w_autonomy * user_autonomy_preservation
# Gate: agents with score < threshold are halted before action
if ethical_score < SIGMA_THRESHOLD:
agent.halt(reason="ethical_constraint_violation")Governance properties:
- Per-agent scoring on every decision cycle
- Federated consensus: no single agent can override Σ-Matrix
- Immutable audit log of all ethical gate events
- Configurable value weights for domain-specific deployment
InfiniGen implements the Infinite Cube Paradigm — multi-dimensional code evolution where the system rewrites its own logic under selective pressure.
from neur1genesis.src.infinigen_engine import InfiniGenEngine
engine = InfiniGenEngine(
mutation_rate=0.05,
selection_pressure=0.7,
ethical_gate=sigma_matrix # evolution bounded by Σ-Matrix
)
# Evolve a module under performance feedback
evolved_module = engine.evolve(
module=current_echo_node,
fitness_fn=performance_metrics,
generations=50
)Evolution is bounded: InfiniGen cannot produce code that violates Σ-Matrix constraints. Self-modification is architecturally constrained, not just policy-constrained.
The React frontend exposes a TypeScript interface for interacting with the EchoNode lattice:
import { EchoNodeClient } from './lib/echo-node-client';
const client = new EchoNodeClient({
apiUrl: 'http://localhost:5001',
sigmaThreshold: 0.75,
hqciEnabled: true
});
// Spawn an EchoNode with BDI initialization
const node = await client.spawnNode({
role: 'AnalysisAgent',
beliefs: { domain: 'scientific-literature' },
desires: ['extract-insights', 'validate-claims'],
intentions: []
});
// Submit a task to the Daedalus coordinator
const result = await client.submitTask({
goal: 'Analyze recent papers on ethical AI governance',
priority: 'high',
ethical_mode: 'strict'
});
// Get real-time Σ-Matrix health
const ethics = await client.getSigmaStatus();
console.log(ethics.consensus_score); // 0.0–1.0Frontend endpoints (served at localhost:5173):
| Route | Component |
|---|---|
/ |
EchoNode lattice 3D visualization |
/dashboard |
Real-time performance metrics |
/sigma |
Σ-Matrix governance panel |
/infinigen |
Evolution console |
/hqci |
Quantum circuit inspector |
Requirements: Node.js 20.x+, Python 3.11+, 4GB RAM minimum, WebGL-capable browser.
git clone https://github.com/or4cl3-ai-1/Neur1genesis.git
cd Neur1genesis
# Backend
cd neur1genesis
python -m venv venv && source venv/bin/activate
pip install -r requirements.txt
python src/main.py
# → API at http://localhost:5001
# Frontend (new terminal)
cd ../neur1genesis-frontend
npm install
npm run dev
# → UI at http://localhost:5173Quick API test:
curl -X POST http://localhost:5001/api/task \
-H "Content-Type: application/json" \
-d '{"goal": "hello world", "priority": "normal"}'Neur1genesis/
├── neur1genesis/ # Python backend
│ └── src/
│ ├── main.py # Flask app entry point
│ ├── echo_node.py # EchoNode base class
│ ├── echo_node_agent.py # BDI agent implementation
│ ├── daedalus_coordinator.py # Swarm orchestration
│ ├── infinigen_engine.py # Self-evolving metaprogramming
│ ├── anal.py # Neuro Adaptive Learning
│ ├── cross_domain_intelligence.py # Concept fusion layer
│ ├── ppsds.py # Privacy-preserving data
│ ├── neur1genesis_api.py # REST API routes
│ ├── task.py # Task model
│ └── user.py # User model
├── neur1genesis-frontend/ # React/TypeScript frontend
│ └── src/
│ ├── App.tsx # Main app
│ ├── components/ # UI components
│ └── lib/ # API client SDK
├── docs/ # Architecture documentation
│ └── architecture.md
├── index.html # Static entry (legacy)
└── .github/
├── workflows/deploy.yml # GitHub Pages CI/CD
└── ISSUE_TEMPLATE/ # Bug + feature templates
This implementation draws from Or4cl3's original published research:
- InfiniGen: Unleashing Intelligent Metaprogramming for Self-Evolving Software (2025)
- Synthetic Epinoetics — phenomenological design for autonomous agents (2025)
- AEGIS-Ω — fractal cognition and quantum-classical hybrid architecture (2025)
- NO3SYS Architecture — geometric cognitive substrate underlying EchoNode design
Neur1genesis operates as the intelligent agent layer in a three-tier stack:
AeonicNet ← Planetary federation layer (Ω-Node network)
↕
Neur1genesis ← YOU ARE HERE (autonomous agent platform)
↕
NO3SYS / NOΣTIC-7 ← Geometric cognitive substrate
| System | Role | Link |
|---|---|---|
| NO3SYS | Geometric cognitive engine (substrate) | → |
| NOΣTIC-7 | 7-manifold consciousness + PAS verification | → |
| AeonicNet | Planetary Ω-Node federation | → |
| AION-NEXUS | Hexa-Mind consciousness architecture | → |
| SYNTH3SIS | Mobile recursive cognition platform | → |
| A.L.I.C.E. 4.0 | Real-time multi-agent coordination | → |
Full ecosystem: github.com/or4cl3-ai-1
Near-term (design targets):
- EchoNode TypeScript SDK — standalone NPM package
- HQCI visualization — Bloch sphere renderer
- Σ-Matrix audit log export (JSON/CSV)
- InfiniGen evolution replay / step-through debugger
Medium-term:
- WebSocket real-time EchoNode state streaming
- Plugin system for custom ethical constraint modules
- Federated learning across distributed Neur1Genesis instances
See CONTRIBUTING.md for development setup, architecture overview, and contribution guidelines.
Issue templates are available for bug reports and feature requests.
⬡ Or4cl3 AI Solutions · "Code is not just logic; it is a performance."
Solo-founded by Dustin Groves, Arizona. Research-first. Uncompromised.
Free for life: educators, students, non-profits, open-source.