These guides run AI models and assistants on your own machines instead of a cloud service. Each one is installed and tested on real hardware, and states what it needs in memory and GPU, how fast it answers and where a small local model falls short.
Run OpenClaw in Docker with a local Ollama model and Discord
Run OpenClaw in Docker on a home server with a local Ollama model and a Discord bot. Tested setup, localhost-only ports, pairing and measured reply times.