The repository identifies Black Forest Labs' FLUX.1-dev as the original model and lists the licence as flux-1-dev-non-commercial-license.
The main checkpoint files are flux1-dev.safetensors and flux1-dev-fp8.safetensors, placed in ComfyUI/models/checkpoints/ and loaded with ComfyUI's Load Checkpoint node. According to the model card, the two Flux text encoders are included in the single safetensor file. It describes the checkpoint as a smaller option for ComfyUI users with less than 24 GB of VRAM.
The repository also lists separate Canny, Depth, Fill and Depth LoRA files, alongside workflows for text-to-image, inpaint, outpaint and reference-image generation. The source does not provide benchmark results, a release date, measured memory use, speed comparisons, or evidence that the repackaging changes FLUX.1-dev's generation quality.