Running open-weight LLMs locally with Ollama: setup, models, hardware and wiring it into self-hosted AI stacks.
Game AI isn't intelligent. It's state machines, search, and scoring arranged to feel like judgment, with real, runnable code from FSMs to local LLM dialogue.
What actually fits on Oracle Cloud free tier after the June 2026 cut: Ghost, Vaultwarden, Syncthing, monitoring, and a small AI model, sized in real RAM.
Building a local AI server is more than installing Ollama. This guide builds a full production AI stack with Open WebUI, LiteLLM, Qdrant, Traefik, and Langfuse — what each component does, why it's needed, and three complete Docker Compose configurations for beginners, power users, and small teams.
Self-host OpenClaw from zero to production. Server provisioning, SSH lockdown, Docker, reverse proxy, Fail2ban, performance tuning, and full backup strategy in one guide.
Build a production-grade self-hosted AI agent from scratch using Docker Compose. Covers the full stack — Ollama, n8n, Flowise, Qdrant, SearXNG, Langfuse, and Redis.
GitHub Copilot charges $19/month and sends your entire codebase to Microsoft's servers. You can build something better — one that knows your whole project, costs $0/month after hardware, and never phones home.
A local LLM is three layers: weights on disk, a runtime that loads them, an interface that talks to it. Once you see the stack that way, the hardware question answers itself and the tuning stops being guesswork.
