NIGHTFRAME

Decentralized AI Mesh Network

NIGHTFRAME is a distributed computing platform designed for collaborative neural network inference across a mesh of autonomous nodes. The system integrates cellular intelligence, GPU-accelerated ONNX runtime environments, and automated network propagation mechanisms.

---

Technical Specifications

Component	Specification
Distributed Inference	Pipeline-parallel model sharding across heterogeneous nodes
Cellular Intelligence	RF fingerprinting and handover prediction using LSTM models
Neural Compute	ONNX Runtime integration with support for CUDA, DirectML, and CoreML
Mesh Discovery	Hybrid discovery utilizing UDP broadcast and mDNS peer exchange
Network Propagation	Platform-aware captive portal for automated node onboarding
Economic Layer	Credit-based incentive system for compute and storage contribution

---

System Architecture

┌─────────────────────────────────────────────────────────────────────────────┐
│                          WEB CONSOLE (Next.js)                              │
└───────────────────────────────────────┬─────────────────────────────────────┘
                                        │ SignalR Telemetry
                                        ▼
┌─────────────────────────────────────────────────────────────────────────────┐
│                         ORCHESTRATOR (Primary Node)                         │
│                         ASP.NET Core / gRPC                                 │
├─────────────────────────────────────────────────────────────────────────────┤
│  Drone Registry │ Ledger Service │ Shard Coordinator │ Cell Coordinator      │
└───────────────────────────────────────┬─────────────────────────────────────┘
                                        │ gRPC Bidirectional Streaming
                        ┌───────────────┼───────────────┐
                        ▼               ▼               ▼
                  ┌─────────┐     ┌─────────┐     ┌─────────┐
                  │ DESKTOP │     │  SCOUT  │     │ ANDROID │
                  │  DRONE  │     │  DRONE  │     │  SCOUT  │
                  │─────────│     │─────────│     │─────────│
                  │ Compute │     │ Cellular│     │ Wi-Fi   │
                  │ Storage │     │ RF Loc  │     │ Scanning│
                  │ Gateway │     │ Predict │     │ GPS Map │
                  └─────────┘     └─────────┘     └─────────┘

---

Navigation and Directory Structure

Path	Description	Documentation
Orchestrator/	Central command node handling gRPC and SignalR	Orchestrator Core
Drone/	Desktop node implementation with ONNX and Cellular modules	Drone Docs
DroneAndroid/	Mobile scout node implemented in .NET MAUI	Mobile Docs
gamma1-web/	Administrative web interface (Next.js)	Web Interface
Shared/	Core interfaces and message contracts	Interface Definitions
Watchdog/	Supervisor process for node reliability and updates	Watchdog Specs
docs/	Technical specifications and setup guides	Documentation Index

---

Technical Deployment

System Requirements

• .NET 8.0 SDK
• Node.js 18 or higher (Web Console)
• Supported OS: Windows 10/11, macOS, Linux (various distributions)

Build Process

1. Clone the repository:

   git clone <repository-url>
   cd Limitless0.1

2. Compile the solution:

   dotnet build NIGHTFRAME.sln

3. Initialize the Orchestrator in Genesis mode:

   dotnet run --project Orchestrator -- --mode genesis

4. Launch a standalone Drone node:

   dotnet run --project Drone

---

Cellular Logic and RF Telemetry

The cellular module implements the following:

• RF Fingerprinting: Localizes nodes without GPS using signal strength patterns (approx. 50-500m precision).
• Handover Prediction: Utilizes LSTM networks to predict optimal tower switching.
• Modem Interfacing: Low-level AT command support for Quectel, Telit, and Sierra Wireless hardware.
• OpenCellID: Synchronizes with global cell tower databases for initial localization.

---

Neural Compute Providers

Provider	Platform Support	Hardware Acceleration
CUDA	Windows / Linux	NVIDIA GPU
DirectML	Windows	Microsoft DirectML compatible GPUs
CoreML	macOS	Apple Silicon / Metal
CPU	All	x64 / ARM64 instructions

---

Security and Integrity

• ECDSA Identity: Nodes are identified via cryptographic key pairs.
• Node Probation: New nodes undergo a "Shadow Mode" evaluation period.
• Consensus Validation: Inference results are verified across multiple nodes where redundancy is required.
• Signed Attribution: Every compute result is cryptographically signed by the providing node.

---

Economic Model (Credit Ledger)

Metric	Credit Impact
Compute Shard Completion	+100 units
Storage Allocation (per GB)	+10 units
Verified Uptime (per hour)	+5 units
Inference Request (Submit)	-50 units

---

Project Status

Platform	Role	Status
Windows	Full Node	Operational
macOS	Full Node	Operational
Linux	Full Node	Operational
Android	Scout Node	Development

---

Compliance and Legal

Copyright © 2024 Malachi Hooper. All rights reserved. Proprietary software.