JustRayzist
A local-first CUDA image-generation app built around the Rayzist Z-Image Turbo fine-tune, with a FastAPI backend, browser UI and CLI, LoRA and wildcard libraries, prompt enhancement and upscaling.
JustRayzist packages a FastAPI backend, a browser-based UI, a Typer command-line interface, launch scripts and model-pack metadata around the Rayzist Z-Image Turbo fine-tune, an image-generation model. After setup it is designed to run entirely offline, and it auto-detects an internal resource tier, high, balanced or constrained, from currently free VRAM to adjust its execution strategy without asking the user to choose a memory mode.
Workflows include gallery management, multi-LoRA support with draft upload and trigger detection, a wildcard library of editable prompt tokens with encoder-assisted suggestions, a documentation-grounded chat assistant, prompt enhancement, and content-aware 2x upscaling with tiled seam repair. It is built on Python, FastAPI, PyTorch, Diffusers and Transformers, with Windows as its primary packaged-install target and Linux and macOS supported through source-mode launcher scripts.