Qwen Image 2.1
Generate and edit images locally on Apple Silicon with Qwen Image 2.1, acceleration LoRAs and Qwen Control.
Mac / MLX guidezimage
FP16 Dense acceleration
An optional computation path for supported models; compare speed and memory use on your Mac.
More native models
FLUX 3, Krea 2 Turbo and Iris join Qwen Image 2.1 and Turbo.
Runs locally on your Mac
Native inference code included; models download separately. Apple Silicon · macOS 15+.
APPLE SILICON · SWIFT / MLX
AI image models for Mac, powered by MLX
Qwen Image 2.1, Qwen Image 2.1 Turbo, FLUX 3, Krea 2 and Iris — native Swift / MLX on Apple Silicon, with optional FP16 Dense acceleration.
Qwen Image 2.1 Turbo
Use the dedicated Qwen Image 2.1 Turbo model locally on Mac with native Swift / MLX inference.
Mac / MLX guideFLUX 3
Run FLUX 3 locally on Apple Silicon with ZImageApp v2.1.0, configurable model paths and optional FP16 Dense.
Mac / MLX guideKrea 2
Generate with Krea 2 Turbo locally on Mac using native Swift / MLX, model path settings and optional FP16 Dense.
Mac / MLX guideIris
Run Iris locally on Apple Silicon in ZImageApp v2.1.0 with native Swift / MLX and optional FP16 Dense.
Mac / MLX guideFP16 Dense
Explore optional FP16 Dense computation for supported image models in ZImageApp on Apple Silicon.
Mac / MLX guideZImageApp v2.1.0
AI image models for Mac, powered by MLX
Qwen Image 2.1, Qwen Image 2.1 Turbo, FLUX 3, Krea 2 and Iris — native Swift / MLX on Apple Silicon, with optional FP16 Dense acceleration.
01
Qwen Image 2.1
Create with Qwen-Image-2.1, then add acceleration and editing LoRAs to shape the result.
- Optional CacheDiT acceleration · off by default
- 4 / 6 / 8 steps acceleration LoRAs
- Face Swap / Head Swap LoRAs
- Try On LoRA / Control
02
Qwen Image 2.1 Turbo
Select the built-in Turbo model and generate locally on Apple Silicon.
- Qwen-Image-2.1-Turbo support
- Built-in q4 quantized model
- CacheDiT off by default · enable when needed
Roadmap
2025/11/30
v0.1
v0.12025/11/30
Initial release with basic ZImageApp generation running locally on Mac.
2025/12/3
v0.2.0
v0.2.02025/12/3
Added Image to Text, Upscale, and Batch Processor features while improving the local creation workflow.
2026/04/24
v1.0.0
v1.0.02026/04/24
Started the 1.0 line with a redesigned UI, separate ComfyUI component loading, Cache DiT speedup, SeedVR2 super-resolution, and Chinese/English UI.
2026/04/25
v1.0.1
v1.0.12026/04/25
Patch release for the signed DMG startup issue in v1.0.0.
2026/10/19
v2.1.0
v2.1.02026/10/19
Adds native FLUX 3, Krea 2 Turbo and Iris, optional FP16 Dense acceleration, model path settings and a standalone signed, notarized app.
Current Release
Coming Soon
Next
NextComing Soon
More quality and performance work for Inpaint, Z-Image Image-to-Image, and Z-Image-Base is planned.
Planned
Coming Soon
Pro / Future
Pro / FutureComing Soon
Images to LoRA, non-Apple-Silicon builds, Mac App Store release, and a one-time-purchase Pro version are planned.
Planned
Images to LoRA Preview
A Mac-friendly Images to LoRA workflow is planned for a future release.
Original
Target Style
ZImageApp v0.2.0 Features
Image understanding, sharper exports, and batch workflows for local creation.
Image to Text
Extract descriptions from images to organize assets and create new prompts.
Upscale
Increase image resolution in one click with clearer details.
Batch Processor
Process large image sets efficiently with fewer repeated steps.
Subscribe for Updates
Get the latest ZImageApp updates, tutorials, and exclusive benefits first.
By subscribing, you agree to receive ZImageApp related news. We respect your privacy.