MOSS-Image

MOSS-Image

Text-to-image and image-edit model files for ComfyUI, repackaged from the OpenMOSS MOSS-Image family. For ComfyUI users who generate or edit pictures in node workflows.

Made by Comfy-Org on Hugging Face. These notes are Mindket's, written from its model page and model card; the model's own files and license are on Hugging Face.

About

What it is: MOSS-Image is a set of model files for ComfyUI that Comfy-Org repackaged from the OpenMOSS MOSS-Image family. It covers the base and SFT image models, an image-edit LoRA, a color and depth control LoRA, text encoders and two VAE files. Ready-made workflow templates are linked from the page.

Who it suits: It suits ComfyUI users who want to make images from text, edit existing pictures or guide a layout in a node workflow.

Industry | Who uses it | How they use it | Example
Design | Designers | Generate pictures from text prompts | Concept visuals for a client brief
Marketing | Content teams | Edit existing images with the edit LoRA | A before and after change on a product photo
Creative tech | ComfyUI builders | Place elements with a layout-guided workflow | A poster with text boxes in set positions
Research | ML hobbyists | Compare the base and SFT checkpoints | Run the same prompt through both models

If you do not use ComfyUI, you can skip it.

Where to use it:

• ComfyUI, with the files placed in its models folders
• The workflow templates from the Comfy-Org workflow templates collection
• Windows, macOS or Linux, wherever your ComfyUI runs

Use case | Who | What they get back
Text to image | Designers | A generated picture from your prompt
Image editing | Content teams | An edited version of an input image
Layout-guided generation | ComfyUI builders | A picture that follows boxes you place

How to set it up: You need these first:

• ComfyUI installed on your computer
• Python, for the Hugging Face download command
• Disk space for the files you choose, as several come in more than one precision

1. Download the model files with the hub tool.
2. Copy the files you need into the matching folders under the ComfyUI models folder, as the table shows.
3. Open one of the MOSS-Image workflow templates in ComfyUI.

pip install -U huggingface_hub
hf download Comfy-Org/MOSS-Image

Title: Where the files go inside the ComfyUI models folder
Folder | Files
diffusion_models | moss_image_base_bf16, moss_image_base_int8_convrot, moss_image_sft_bf16, moss_image_sft_int8_convrot
loras | moss_image_color_depth_control, moss_image_edit_lora_rank_64_bf16
text_encoders | qwen3.5_9b_bf16, qwen3.5_9b_int8_convrot, qwen3.5_9b_w4a8_gptq
vae | flux2-vae_small, flux2-vae

How to use it: Open a workflow template in ComfyUI, check that its model nodes point at the files you placed, type a prompt or load an image, and run it.

Tool | What it does
MOSS-Image: Text to Image | Makes a picture from a text prompt
MOSS-Image Base: Text to Image | Makes a picture from a prompt with the base model
MOSS-Image: Image Edit LoRA | Edits an input image, and the page shows a before and an after
MOSS-Image: Layout-Guided Text to Image | Makes a picture from a prompt plus a layout of boxes

Example: You open the Text to Image template, type a short description of a scene, and run it. What comes back is a new picture that follows your description, saved by ComfyUI like any other output.

For newcomers and experts: This is a pack of files, not an app, and you use it inside ComfyUI. A good first step is to place one diffusion model, one text encoder and one VAE, then open the Text to Image template.

The file names show that the diffusion models and the Qwen3.5 9B text encoder come in several precisions, including bf16, int8 and a w4a8 gptq encoder. The edit LoRA is named as rank 64. The size and quality trade-off between precisions is yours to test.

Place the model files in ComfyUI folders
Open a MOSS-Image workflow template
Add a prompt, an image or a layout
Run it to get a picture

Good to know: The page is a repackaging for ComfyUI, so the original model details sit in the OpenMOSS repositories it names.

• Sources: the files come from the MOSS-Image-Base, MOSS-Image-SFT, color-depth-control and Moss-Image-Edit repositories
• Other versions: 0 quantized and 2 fine-tuned versions made by other people are listed
• License: Apache 2.0
• Cost to run: none named
• Data it reaches: the prompts and images you give the workflow, and the model files in your ComfyUI folders

On Hugging Face, 13 November 2026

• Library: diffusion-single-file
• Downloads in the last 30 days: 6,310,763
• Likes: 666
• License: apache-2.0
• Based on: OpenMOSS-Team/MOSS-Image-Base
• Fine-tunes: 2
• Tags: diffusion-single-file, comfyui

Links

• Model: https://huggingface.co/Comfy-Org/MOSS-Image
• Files: https://huggingface.co/Comfy-Org/MOSS-Image/tree/main

Get the model

Download its files with Hugging Face's command line:

pip install -U huggingface_hub
hf download Comfy-Org/MOSS-Image

Then follow the usage in its model card. Nothing in this kit runs by itself.