- Posted on
- Featured Image
Guide to building an AI homelab that runs private LLMs on Debian/Ubuntu, Fedora/RHEL, or openSUSE, with Bash-first steps for CPU and GPU. Covers installing tools and containers, fastest path via Ollama (local REST API), DIY via llama.cpp, optional NVIDIA/AMD acceleration, serving and securing services, firewall and systemd, plus tips on models, storage, and monitoring—so you can stand up a local API fast and scale confidently.