Loop Forge

A cafe front from layout boxes

From MOSS-Image add-ons, tested locally

Frame from the The boxes result
The boxes
Frame from the Layout result result
Layout resultDownload JPG0.2 MB
Frame from the Boxes drawn on the output result
Boxes drawn on the output
Model
MOSS-Image SFT (moss_image_sft_int8_convrot.safetensors)
Output
1024×1024
Sampling
25 steps, CFG 2.5, er_sde / simple
Seed
101
Layout result
43.0 s, peak VRAM 10.7 GiB
Ran on
RTX 3060, 12GB (local)
Verdict

The sign and the chalkboard are spelled right, and he is standing about where I drew his box.

The prompts

Download workflow
Warm photograph of the front of a small corner cafe at golden hour: a young man stands by the open door holding a takeaway coffee cup, a dark green wooden shop sign above, a standing chalkboard beside him, a small counter with a steaming cup and croissants in the window. Soft warm light, shallow depth of field, 35mm photo.

Notes

The boxes are drawn in the template's Create Bounding Boxes node. Each one is a text box (the exact words to spell and how they should look) or an object box (a description), and the MOSS-Image Layout Prompt node turns the caption and the boxes into the layout prompt the model was trained on. The Caption and Boxes tabs are exactly what those two nodes were given.

Box 2 is a character box: the man is described in words. The layout template takes no images.

Comfy-Org workflow templates (the MOSS-Image templates ship with ComfyUI)

Run it yourself

This test loads no input files. Put the models below in place and drag the workflow onto ComfyUI.

FilePut it inSize
moss_image_sft_int8_convrot.safetensorsmodels/diffusion_models5.4 GB
qwen3.5_9b_w4a8_gptq.safetensorsmodels/text_encoders7.3 GB
flux2-vae_small.safetensorsmodels/vae249.5 MB

No custom nodes. Every node is part of ComfyUI 0.43.1 (frontend 1.57.6, templates 0.11.100).