Local AI runners are proliferating fast on Mac

Nativ launched today as a native Mac app for running frontier open models locally, and the HN thread immediately turned into a comparison grid: Ollama, LM Studio, Bionic, rapid-mlx, llama.app. One commenter had explicitly called for more Ollama competitors in a previous thread and noted that the call had been answered. The space is filling in fast.

The pattern: Apple Silicon's MLX framework has made local inference genuinely viable on consumer hardware, and now the tooling layer is catching up. A year ago Ollama had this category mostly to itself. Now there are at least five named options with active development, and the differentiation is shifting toward UI quality, model management, and integration with local workflows rather than raw inference capability.

Kimi Work, also discussed today, is a Chinese AI desktop app that commenters noted is more of a chat interface than a true desktop environment. The appetite for a real AI-native desktop OS is clearly there, but nobody has shipped it yet.