AI-MC2 Fabric Intelligence Layer SDK

Python SDK for the AI-MC2 autonomous intelligence layer for multi-cluster congestion control. Zero dependencies — standard library only.

Install

# Open-source SDK (MIT) — zero dependencies
pip install ai-mc2

# From source (clone + editable install)
git clone https://github.com/HONEYPOTZ-AI/ai-mc2.git
cd ai-mc2
pip install -e .

Quick Start

from ai_mc2.topology import LatencyModel
from ai_mc2.models import TopologyType, FlowPriority

# Compare AI-MC2 vs Fat-tree latency at 6 hops
model = LatencyModel(TopologyType.AIMC2)
latency = model.p99_us(6, FlowPriority.LATENCY)
print(f"AI-MC2 p99 @ 6 hops: {latency:.2f} µs")  # 2.89 µs

# AI-MC2 is ~8× faster than Fat-tree
c = model.compare_all(6)
print(f"Fat-tree: {c['fat_tree']:.1f} µs, AI-MC2: {c['aimc2']:.1f} µs")

# CLI demo
python scripts/demo.py

REST API (Proprietary)

# Install API extras
pip install ai-mc2[api]

# Start the API server
uvicorn api.main:app --reload --host 0.0.0.0 --port 8080

# Open OpenAPI docs
open http://localhost:8080/docs

Subsystems

Development

make install-dev   # Install SDK + API + test deps
make test          # Run unit + integration tests
make demo          # End-to-end walkthrough
make kpi           # Validate KPI targets
make smoke         # Smoke-test running API
make serve         # Start dev server on :8080

Deployment

# Docker
docker compose up -d

# Kubernetes
kubectl apply -f deploy/k8s/all.yaml

# Helm
helm upgrade --install ai-mc2 deploy/helm/ai-mc2 \
  --set ingress.host=ai-mc2-api.your-domain.com

License & Pricing

This repository uses a dual-tier license model:

SDK is open-source (MIT) to drive adoption and build trust — the moat is in the operational, deployment, and compliance tooling.

📧 mc2@honeypotz.net for Pro/Enterprise inquiries.