A1AI is a homelab-first blog about running real AI infrastructure on hardware you own — no subscriptions, no vendor lock-in, no cloud bill.
What We Cover
- Local LLMs — running Ollama, LiteLLM, and open-weight models on your own GPU
- Self-hosted tooling — n8n, Ghost, Supabase, Listmonk, and the full open-source stack
- Homelab builds — hardware, networking, Tailscale, Docker, and real-world configs
- AI automation — workflows that actually run in production without SaaS dependencies
The Setup
Everything on A1AI runs on a 4-node homelab:
- Moolah — RTX 5070, 47GB RAM — Ollama, LiteLLM, Supabase, ComfyUI
- Neo — Ghost CMS, n8n, Home Assistant, Caddy
- Uno — Listmonk, Pihole, Uptime Kuma
- VPS — Cloudflare tunnel gateway, no home IP exposed
All connected over Tailscale. Zero cloud dependencies for core services.
The Daily Brief
Every morning at 8am, an n8n workflow pulls from 5 AI news sources, runs each article through a local LLM (llama3.1:8b), and publishes a structured A1AI briefing automatically. No human in the loop. No API costs. Starter members get the full workflow file.
Work With Me
I offer homelab architecture audits for teams or individuals who want to move off cloud AI APIs and run their own stack. If you are spending real money on LLM APIs and want to explore local inference for your use case, get in touch.
Join the Newsletter
The A1AI Newsletter covers new builds, workflow templates, and the daily AI brief. Free to subscribe.