HaDeS MOSS Workflows

This page gathers my ComfyUI workflows for MOSS-Image, each one a new version here. They are built around the HaDeS MOSS Suite nodes and kept simple enough to load and run. Every setting in them comes from measurements, not guesswork, and the measurements are public in the suite's tools/RESULTS.md.

More workflows will follow (merging, LoRA tools, layout prompts...). Follow the page to get them as they come out.

Simple Turbo 8 steps + Upscale x2

A simple MOSS-Image workflow: generate at 8 steps with our turbo LoRA, then optionally double the size.

Each step is a group that you switch on or off with the Fast Group Bypasser. Groups 2 and 3 are off by default, so you only render the 1024 image until you turn one of them on.

The turbo LoRA used here: HaDeS MOSS Turbo 8-steps v0.7 → [Civitai link to the Turbo model page], (also on Hugging Face: https://huggingface.co/DeVilDoNia/HaDeS-MOSS-Turbo).

The three groups

1. Generate (always on): MOSS-Image SFT + HaDeS Turbo 8-steps v0.7, 8 steps, CFG 1, up to 1024 px.

2. Upscale x2 (model): RealESRGAN x2 doubles the image in pixels in about a second. It doesn't redraw anything.

3. Upscale x2 + MOSS pass: re-renders the upscaled image at denoise 0.25 with the same turbo. It adds hair, pores and fine texture that the upscaler smooths away. Group 3 needs group 2 on.

What you need

- HaDeS MOSS Suite (ComfyUI Manager or https://github.com/Devildonia/HaDeS-MOSS-Suite)

- rgthree-comfy for the Fast Group Bypasser

- Models:

- diffusion_models: moss_image_sft_bf16 (or int8_convrot)

- text_encoders: qwen3.5_9b_int8_convrot

- vae: flux2-vae

- loras: HaDeS-MOSS-Turbo-8steps-v0.7

- upscale_models: RealESRGAN_x2plus.pth

Without the turbo

Unplug the turbo input of the MOSS KSampler. It then runs SFT at 25 steps and CFG 2.5.

The measurements are in tools/RESULTS.md of the HaDeS MOSS Suite repo.