Hugging Face announced that Jun Kim, the creator and maintainer of oMLX, has joined the company full-time. oMLX is a community-driven ecosystem built on top of Apple's MLX framework, designed to make running large language models on Apple Silicon fast and accessible to a broader audience of developers and researchers.
The move signals Hugging Face's deepening commitment to local AI inference as a viable alternative to cloud-only deployments. With MLX optimized for Apple's unified memory architecture, the framework has become a go-to choice for developers who want to run models like Llama, Mistral, and other open-weight architectures on Macs without requiring expensive cloud infrastructure or GPU clusters.
For oMLX specifically, the transition from a side project maintained in Kim's spare time to a fully funded effort within Hugging Face means faster development cycles, better documentation, and long-term stability for the growing user base. The project remains Apache 2.0 licensed, and Kim will continue leading it with the same open-source philosophy that guided its creation.
Hugging Face says the partnership aims to streamline the path from a transformers model definition to a reference MLX implementation that multiple inference engines can consume. The collaboration extends to mlx-lm, mlx-vlm, LMStudio, and other projects in the local-AI stack. The goal is to make the Hugging Face Hub the central discovery point for MLX-converted models while upstreaming improvements wherever they benefit the broader ecosystem of local inference tooling.