Edge AI Infrastructure

Platform for Nexus Neural Network Nodes

Bootstrap complete edge AI systems in minutes. Unified platform combining MING stack, Edge Impulse, and Gen AI layer for IoT engineers building distributed intelligence.

Get Started → View Architecture
Concept

Nexus Neural Network Nodes (N4)

A distributed edge intelligence architecture where each node runs on-device ML inference while participating in a coordinated mesh network. Data flows from devices through orchestration layers to storage and visualization, with AI agents providing automation and natural language interfaces.

For Developers

Rapid Deployment

Quick Start


git clone [email protected]:raisga/p4n4-docker.git
cd p4n4-docker
docker compose up -d

# Services ready:
# MQTT: localhost:1883
# Node-RED: localhost:1880
# InfluxDB: localhost:8086
# Grafana: localhost:3000
# n8n: localhost:5678
# Ollama: localhost:11434

Architecture Benefits

  • Containerized services with Docker Compose
  • Pre-configured networking and authentication
  • Persistent volumes for data retention
  • Environment variable configuration
  • Extensible service definitions
  • Production-grade logging and monitoring

Open Source Infrastructure

P4N4 is built on open-source foundations. Community-driven development, transparent roadmap, MIT licensed. Contributions welcome.

View on GitHub → Join Community
Architecture

System Design

Three-layer architecture: device connectivity, data orchestration, and AI automation. Each component is containerized and can be deployed independently or as a complete stack.

Device Layer
IoT Sensors
Edge Devices
Robotics Nodes
Industrial Equipment
↓
Edge AI
Edge Impulse (ML Models)
On-device Inference
↓
MING Stack
MQTT (Messaging)
InfluxDB (Storage)
Node-RED (Orchestration)
Grafana (Visualization)
⇕
Gen AI Layer
n8n (Workflows)
Letta (Agent Memory)
Ollama (Local LLM)
Features

Built for Edge AI Development

01

Unified Bootstrap

Single Docker Compose deployment. All services pre-configured and networked. Start building in minutes, not days.

02

Edge-Native Pipelines

Purpose-built for edge AI workflows. Device messaging, data orchestration, ML inference, and monitoring in one stack.

03

Local-First AI

Run LLMs locally with Ollama. No external API calls. Complete data sovereignty and offline capability.

04

Modular Stack

Use what you need. Each component can be deployed independently or as part of the complete platform.

05

Developer-First

Built by engineers, for engineers. Clear APIs, comprehensive docs, and extensible architecture.

06

Production Ready

Battle-tested components. Industry-standard protocols. Scales from prototype to production.

Applications

Built For Real-World Systems

Smart Sensors

Deploy ML-powered sensor networks. Process environmental data, detect anomalies, trigger automated responses.

Warehouse Safety

Real-time edge AI monitoring system for warehouse safety compliance and loss prevention.

AI-Controlled Systems

Let agents make decisions about physical systems. Natural language queries over sensor data and device control.