Setting up OpenCode with Ollama and sbx on Mac
Local LLM development is now viable using Ollama and OpenCode on Mac, with powerful models like Qwen 3.8 and Gemma4 that can run efficiently on Apple Silicon. The author provides a setup guide for configuring these tools in Docker sandboxes (sbx), including installation steps via Homebrew and model downloads, along with configuration files needed to integrate Ollama with OpenCode for local web application development. The recommended hardware is a MacBook Pro with at least 48GB of memory to run larger 30B+ models with adequate context windows.