In the last few months, there has been rapid development in the world of local LLMs, which can run in PCs and laptops with enough RAM and a GPU. Though still far behind from frontier models, local LLMs can still write code, transcribe handwritten notes to LaTeX, and perform other agentic tasks without sending any data to a third party. In this talk we will explore the wide range of local LLMs that exist, what sets them apart from each other and areas where they are still lacking. We will also learn how to run these models, set up your own workflows and evaluate their outputs.