MiniMax-H3 Runs on Apple Silicon Now, Thanks to an MLX Port
minimax-h3-mlx is a Python package that ports MiniMax-H3, an omni-modal model accepting text, image, audio, and video input, to Apple's MLX framework for running locally on Apple Silicon.
What it is
A community MLX port of MiniMax-H3, the omni-modal generative model MiniMax released days earlier, packaged as the PipeNetwork/minimax-h3-mlx repository.
What it does
The model accepts text, images, audio, and video, and can generate video clips up to 15 seconds long with audio included; the MLX port lets it run on Apple Silicon instead of requiring a cloud GPU or an Nvidia setup, and Simon Willison confirmed it running on an M5 Max MacBook Pro.
Why it matters
An omni-modal video generation model running locally on a laptop, instead of behind an API, is a meaningfully lower bar for experimenting with this class of model, similar in spirit to Nativ bringing MLX chat models to a local Mac app.
How to use it
Clone the minimax-h3-mlx repo on an Apple Silicon Mac and follow its run instructions to generate video clips from text, image, or audio input locally.
condense-json 1.0: One Library, Eighteen Months, Zero Drama
condense-json 1.0 is a Python library that compresses deeply nested JSON into a shorter, still-readable form for pasting into LLM prompts or logs.
datasette-agent 0.4a0 Lets Agent Tools Run JavaScript in Your Browser
Datasette Agent 0.4a0 adds a browser_task mechanism that lets agent plugin tools execute custom JavaScript directly in the user's browser instead of on the server.
Nativ Puts a Local Chat UI and API Server on Every MLX Model on Your Mac
Nativ is a macOS desktop app from MLX-VLM developer Prince Canuma that wraps MLX with a chat interface and a localhost API server for running AI models locally on Apple Silicon.