Wanwei Yiyan Smart Emoji Design

Problem

When you already have a character illustration, pet photo, or IP image and want to turn it into a usable chat-sticker set, the hard part is not generating one image. It is keeping facial features, color, and line style consistent, producing a grid layout that can be split cleanly, and exposing a task result that can be previewed and downloaded.

How It Works

The skill builds an asynchronous workflow around prompt and reference_image. It validates input, parses the reference image for character traits, art style, and palette, then fuses those signals with the grid setting to generate drawing prompts and slicing parameters. In real execution, execute_task.py submits the job and returns only a task_id; query_task.py polls the job state and syncs each round’s status until the final payload contains original_image and an emoji_list of individual sticker URLs. It supports grid_num=9 or grid_num=16, which fits portrait stickers, cartoon IP expansion, and pet sticker sets.

Boundaries

The flow is asynchronous and should not be treated as a synchronous image endpoint. reference_image should be clear and contain a complete subject, while prompt should describe emotion, style, and usage context. Without showapi_appkey, the run returns a simulated preview result rather than a real generated asset.