Run It

Running models day to day - Ollama, Open WebUI, RAG, coding assistants, and everything else once your setup is live.

Installing Open WebUI: A Proper Chat Interface

Install Open WebUI so you get a real, ChatGPT-style interface for your local models instead of a bare terminal.

15 min read

Docker Compose for a Maintainable Open WebUI Setup

Move from one long docker run command to a single reviewable file - the bridge from a beginner setup to one you can actually maintain.

9 min read

Choosing Your First Model

Parameter count, quantization, and picking a model that actually matches your hardware instead of just grabbing the biggest name you've heard of.

14 min read

Ollama vs. llama.cpp: Why I Still Recommend Ollama

A real debate happened about whether Ollama is worth using at all. Here is what the criticism actually gets right, and why this site still defaults to it.

17 min read

Talking to Your Own Documents with RAG

Set up retrieval-augmented generation in Open WebUI so your local model can answer questions using your own files, not just what it was trained on.

15 min read

Using a Local Model as a Coding Assistant

Connect Ollama to VS Code with the Continue extension so you get AI coding help that never sends your code anywhere.

15 min read

Turning Your Local Model Into an Always-On Bot with OpenClaw

Connect OpenClaw, a self-hosted AI agent framework, to your local Ollama models so you get a persistent assistant in Discord or Telegram, not just a browser tab.

14 min read

Keeping Everything Updated

Updating Ollama, your models, Open WebUI, and OpenClaw safely - what to run, in what order, and what not to do to your Docker volumes.

10 min read

Generating Images Locally

Stable Diffusion, running entirely on your own hardware - what tool to start with, what VRAM actually gets you, and where new images and models come from.

14 min read

Automating Tasks with the Ollama API

Ollama isn't just a chat interface backend - it's a REST API you can script against directly, from a shell command or a few lines of Python.

13 min read

Giving Your Local AI Real Tools with MCP

Model Context Protocol lets your local model read files, fetch web pages, and query databases - not just answer from training data.

16 min read