Local LLM Hosting

Every agent in the main course calls a hosted API, Claude or Gemini, and for almost everyone, that's the right call: better reasoning, no hardware to manage, nothing to babysit. This short, optional course covers the exception: the specific situations where running a model on your own machine actually makes sense, and the fastest way to do it when it does.

It's deliberately small. If you're looking for deep infrastructure, quantization formats, serving engine internals, GPU memory tuning, that's not what this is; it's the practical 20% that covers why and how, for the cases where local inference is genuinely the right tool.

What's here

• Running Models Locally , when local inference is actually worth it, and getting a model running with Ollama in minutes.
• Ollama Cheat Sheet , the commands you'll actually reach for day to day.

Most of the time, you want AI Agents & Vibe Coding instead, this course is here for the specific cases where a hosted API isn't the right fit.