I built a native macOS app for AI image generation, originally for my own creative work.

I enjoy working with models like Z-Image and FLUX, but many existing tools feel closer to research interfaces than creative tools. They expose a lot of power, but often interrupt creative flow.

This project is built around a creator-first workflow: - Easy to get started, with no configuration beyond downloading or importing models - Progressive controls: advanced options are available, but not required - A built-in prompt library to save, reuse, and evolve prompts over time - Lightweight prompt refinement that learns from your saved prompts - Simple add-ons for upscaling and background removal

Under the hood: - Runs Stable Diffusion, FLUX, and Z-Image fully locally - Optimized for Apple Silicon using MLX and Metal

This is an opinionated attempt to make local image generation feel more like a creative tool and less like a technical setup.