What is the easiest way to run AI image models locally on a Mac?

Install ZImageApp on an Apple Silicon Mac with macOS 15+, choose a model such as Qwen Image 2.1 or FLUX 3, download its MLX package in the app and generate. No Python or terminal setup is needed.

Which Macs can run these models?

ZImageApp requires an Apple Silicon Mac (M-series, arm64) running macOS 15 or later. Intel Macs are not supported. Available unified memory limits the resolution and computation mode you can use.

Do I need Python, ComfyUI or the terminal?

No. ZImageApp is a native Swift / MLX Mac app. Install the DMG, choose a model, download or select a compatible MLX package and generate — no Python environment, ComfyUI workflow or command line setup.

Yes. ZImageApp is a free download for Apple Silicon Macs. The installer is Developer ID signed and Apple notarized.

Are my prompts and images uploaded?

No. Generation runs locally on your Mac. A network connection is only needed to download the app and model files.

ZImageApp v2.1.0 · Apple Silicon · macOS 15+

How to run Qwen Image, FLUX 3, Krea 2 and Iris locally on Mac

The simplest way to run Qwen Image 2.1, Qwen Image 2.1 Turbo, FLUX 3, Krea 2 Turbo and Iris Turbo locally on an Apple Silicon Mac: install the free ZImageApp, pick a model and generate offline with native MLX — no Python required.

Free download v2.1.0 · Apple Silicon

Developer ID signed and Apple notarized. Models download separately.

Why run image models with ZImageApp

ZImageApp is a free native Mac app that runs open image generation models on Apple Silicon with Swift and Apple MLX. It supports Qwen Image 2.1, Qwen Image 2.1 Turbo, FLUX 3, Krea 2 Turbo and Iris Turbo in one interface, and earlier releases added Z-Image-Turbo workflows.

Instead of setting up Python, PyTorch or ComfyUI, you install a signed and notarized DMG, select a model and let the app download a compatible MLX package — or point Settings → Models to a package you already have. Prompts and images are processed on your Mac; supported models also offer an optional FP16 Dense computation path.

How to get started

Download the free ZImageApp DMG for Apple Silicon (macOS 15 or later) and move it to Applications.
Open the app and choose a model: Qwen Image 2.1, Qwen Image 2.1 Turbo, FLUX 3, Krea 2 Turbo or Iris Turbo.
Download the compatible MLX model package in the app, or select an existing one in Settings → Models.
Enter a prompt and generate locally; optionally compare FP16 Dense with the default mode.

FAQ

What is the easiest way to run AI image models locally on a Mac?: Install ZImageApp on an Apple Silicon Mac with macOS 15+, choose a model such as Qwen Image 2.1 or FLUX 3, download its MLX package in the app and generate. No Python or terminal setup is needed.
Which Macs can run these models?: ZImageApp requires an Apple Silicon Mac (M-series, arm64) running macOS 15 or later. Intel Macs are not supported. Available unified memory limits the resolution and computation mode you can use.
Do I need Python, ComfyUI or the terminal?: No. ZImageApp is a native Swift / MLX Mac app. Install the DMG, choose a model, download or select a compatible MLX package and generate — no Python environment, ComfyUI workflow or command line setup.
Is ZImageApp free?: Yes. ZImageApp is a free download for Apple Silicon Macs. The installer is Developer ID signed and Apple notarized.
Are my prompts and images uploaded?: No. Generation runs locally on your Mac. A network connection is only needed to download the app and model files.