Can Small Local LLMs Code? Testing LM Studio with OpenCode

In this video, I test whether a relatively small *local LLM* can write working code in an agent-based workflow. I set up LM Studio to run a local language model and connected it with OpenCode, allowing the model to act as a coding agent: reading files, making changes, and generating code directly in the project. For this first experiment, I keep the task simple: building a basic HTML todo application. The goal is to see whether a small local model can successfully follow instructions, produce functional code, and behave like a useful coding assistant without relying on a cloud-based LLM. This is the first video in a mini-series exploring how local LLMs can be used for coding, automation, and eventually bioinformatics workflows. Tools used: LM Studio OpenCode Topics covered: Running a local LLM Setting up an agent-based coding workflow Using OpenCode with LM Studio Testing whether small LLMs can generate working code Building a simple todo app as a first coding experiment