In the Skill's Own Terms
A four-sticker pipeline that turns a user photo — a person, pet, object or logo — into four animated GIF stickers in Funko Pop / Pop Mart blind box style, rendered to look like C4D / Octane 3D figurines on a white studio background with black-outlined captions at the bottom. The four defaults are happy waving (hi), laughing hard (laugh), crying tears (cry) and heart gesture (love).
Captions can be user-supplied, 1–3 words, or pulled from a multilingual captions table matched to the user's detected language. It runs three scripted stages — MiniMax image generation, image-to-video animation, and mp4-to-GIF conversion — and finishes with a deliver_assets block.
What it produces
- Four animated GIF stickers in Funko Pop / Pop Mart style, the declared output format and count.
- Four static 1:1 sticker images written to the output folder, one per default action.
- Four intermediate MP4 clips, one per sticker, produced from the static images.
- A final deliver_assets block listing the four GIF paths, placed last in the response.
How It Works
- 01Settle the Captions
The agent asks in the user's own language whether captions should be custom or default.
- 02Generate Four Static Stickers
The photo is classified as person, animal, object or logo, then four 1:1 images are generated from the image prompt template with {action} and {caption} filled in.
- 03Animate Each Image into Video
Each PNG goes through the video script in image-to-video mode using the video prompt template, producing one MP4 per sticker.
- 04Convert Videos to GIF
A single call to the conversion script takes all four MP4s and writes a GIF next to each one.
- 05Deliver in a Fixed Order
The reply is a one-line status, then the deliver_assets block listing the four GIFs, with nothing written after it.