Describe a look in one sentence, use text-to-image to generate an avatar, then use reference-image editing to produce a character-consistent sticker pack.
Disclaimer: All content shown on this page is generated by AI models and is for reference only; it does not constitute any guarantee of results. Actual output and Credits consumption are subject to "Usage Details" in the console.
wan2.7-image text-to-image (avatar) → wan2.7-image reference-image editing (sticker pack, keeping the character consistent).
Result
Running the "round-faced orange cat" character design through the full pipeline produces this IP avatar and four-panel sticker pack:
 |  |
Step 1: Text-to-image to produce the avatar
First, pin down the look with one sentence (template: species + personality + color scheme + art style):
wan2.7-image at size 1024*1024:
Step 2: Reference-image editing to produce the sticker pack
Pass the avatar from the previous step as a reference image to wan2.7-image to generate a four-panel sticker pack:
Notes
- Consumption: This playbook involves image generation (
wan2.7-image) twice—once as pure text-to-image, once with reference-image editing. - Cost difference: With reference-image editing, the reference image input is the biggest cost driver. To keep the character consistent, this step is noticeably more expensive than pure text-to-image.