











Create stunning AI images locally with Z-Image, Flux.2, ERNIE-Image and QWEN-Image models — fast, private, and optimized for Apple Silicon.
Harness the absolute best in open-source AI models. Deliver unmatched structural accuracy, rapid base generation, and hyper-realistic upscaling at your fingertips.
Available in 9B and 4B variants. Achieve higher fidelity in just 4 steps, with native advanced image editing.
SOTA • NewPremium 8-step base generation delivering uncompromising quality tailored for standard creative workflows.
Core • FastNext-generation vision-language model. Exceptional at understanding complex prompts and generating high-fidelity visual content.
SOTASchnell & Dev models alongside Kontext. A proven standard for pristine visual realism and structural accuracy.
Custom style tuning, character consistency, and fine-grained visual control.
Robust, precise subject isolation and instant transparent backgrounds.
High-resolution upsampling that preserves intricate details and enhances textures.
Optimized for Apple Silicon chips
MindCraft Studio is designed for high-performance local inference with absolute privacy. While we now offer optional Cloud Models for heavy tasks, our core experience remains powered by the MLX framework, running directly on your Mac.
Got an idea or opportunity? We'd love to hear from you. Drop us a message and let's keep in touch.
Alternatively, you can also submit a feature request here.
MindCraft Studio is a professional local-first AI image generation application designed specifically for macOS. It leverages the Apple MLX framework and Metal acceleration to provide industry-leading performance on Apple Silicon (M1/M2/M3/M4) chips. MindCraft offers an infinite canvas for spatial organization and supports state-of-the-art models including FLUX.1, FLUX.2 Klein, and Z-image turbo, while also seamlessly integrating cloud-based powerhouses like ERNIE-Image, ChatGPT-Image-2, Qwen-Image-2, Google Nano Banana, and SeeDream 5.0. All while maintaining 100% user privacy by running offline by default.