🚀 FREE STARTER PACK — workflow + quick-start + cheat sheet, free on Whop: https://whop.com/imagelab/free-starter-moss-image-text-to-image-workflow/

💰 FULL SYSTEM — MOSS-Image Fast Workflows, $39 one-time, code MOSS30 = 30% off ($27.30): https://whop.com/imagelab/moss-image-fast-workflows-text-to-image-edit-layout-boxes-comfyui/

FREE vs FULL: this download is the text-to-image workflow only. The full pack adds the Edit workflow, Layout Boxes workflow, Turbo 4-step workflow, the full prompt library and EN+AR guides.

MOSS-Image text-to-image in ComfyUI — a fast, few-step workflow built from the official template.

Settings: MOSS-Image checkpoint (moss_image_sft_int8_convrot.safetensors), Qwen 3.5 9B text encoder (qwen3.5_9b_w4a8_gptq.safetensors), FLUX2 small VAE (flux2-vae_small.safetensors); 1024×1024, 25 steps, CFG 2.5, er_sde sampler, simple scheduler. The built-in prompt upsampler rewrites a short prompt into one full scene caption before generation. Strong bilingual text rendering and automatic layout. Negative prompt: nsfw, worst quality, bad anatomy, watermark, jpeg artifact.

Needs a real GPU — if VRAM is tight, lower the resolution or use CPU offload. Free official template: import the JSON in ComfyUI and run.

The full MOSS-Image pack — T2I, edit and layout-box workflows plus a settings library and quick-start guide — is on Whop: https://whop.com/imagelab/moss-image-fast-workflows-text-to-image-edit-layout-boxes-comfyui/