System Architecture

A private, production-ready AI infrastructure built on Apple Silicon.

Overview

Two Mac Studio M4 Max nodes with 36GB unified memory each, connected via 10GbE Ethernet for distributed AI workloads.

Connection Details

Parameter	Value
Inter-node Link	10GbE Ethernet
Power	Separate UPS protection
Network	Dedicated VLAN

Node 1 Specifications

• Mac Studio M4 Max
• 36GB Unified Memory
• OpenClaw orchestration layer
• Ollama MLX for LLM inference

Node 2 Specifications

• Mac Studio M4 Max
• 36GB Unified Memory
• ComfyUI for image generation
• Text Gen WebUI for roleplay