A nine-node ComfyUI workflow for MOSS Image. No custom nodes, no LoRA, no ControlNet — it even runs on an 8 GB laptop GPU.WHAT YOU NEED (3 files, all from the MOSS base model)- models/diffusion_models/ -&gt; MOSS base int8- models/text_encoders/ -&gt; Qwen3.5 9B int8- models/vae/ -&gt; flux2-vaeTHE FOUR NUMBERS THAT MATTERSteps 50 · CFG 3.0 · euler / simple · Shift 5.0Sizes: 1024x1408 portrait · 1408x1024 landscape · 1024x1024 square.PROMPT STYLE — Danbooru tags, comma separatedmasterpiece, best quality, very aesthetic, safe, 1girl, solo, &lt;hair/eyes/outfit&gt;, &lt;pose&gt;, &lt;camera&gt;, &lt;scene&gt;, &lt;light&gt;THREE THINGS THAT SURPRISE PEOPLE- It shortens long garments — write "floor length, long skirt"- It drops hoods — write "hood up, head covered"- "tiny figure" can erase the subject — use "full body, medium distance"SPEED ON 8 GB 3070TI ~150 s at 1024x1408, ~130 s at 1408x1024, ~70 s at 1536x512.IF VAEDECODE FAILS with comfy_kitchen / group_norm_stats3dUpdate the kernel library inside ComfyUI's Python (python_embeded\python.exe -m pip install -U comfy-kitchen, needs 0.2.40+) and restart.

MOSS Image — Simple Text2Image (8GB VRAM)

Initial release. Nine-node workflow, no custom nodes. Verified at 50 steps / CFG 3.0 / euler-simple / shift 5.0 on an 8 GB laptop GPU (~150 s per 1024x1408 image).

Files with SHA256: 2e8108be45d9d8b9ac5021701a270aa06215056038755fb925f9019aee964828