# Mine StableDiffusion

A Kotlin Multiplatform client for running Stable Diffusion image models such as SDXL, FLUX, and SD3 offline on Android, iOS, and desktop, built on the stable-diffusion.cpp engine.

Mine StableDiffusion is an offline-first Stable Diffusion client built with Kotlin Multiplatform and Compose Multiplatform, sharing one interface across Android, iOS, and desktop. It runs entirely on-device through a C++ backend with JNI bindings, so generation and any prompts stay local.

It generates images from text, batches multiple variations of a prompt, and supports dragging in external LoRA files in the safetensors format with weight sliders to mix several at once. Users can switch between samplers such as Euler a, DPM++ 2M, LCM, and TCD, and control settings like flash attention, CPU offloading, and direct convolution to fit available VRAM.

Exported PNGs embed their generation parameters, including prompt, seed, model, sampler, and LoRAs, so an image can be reopened in a Stable Diffusion tool to recover its exact setup.

Website: https://github.com/Onion99/KMP-MineStableDiffusion