GuideAnimated Sticker StudioIntermediateAbout 20 min5 steps
Turn one character image into a full pack of animated chat stickers
This guide is for creators and small-business owners who are comfortable running a few terminal commands. You install Animated Sticker Studio, give your agent a character, and get an animated sticker pack back.
Before you start
- Claude Code or Codex, already installed.
- Python 3 with the Pillow library.
- ffmpeg, needed only for the Telegram export or the MP4 preview.
- An image-generation route: Codex has a built-in image tool. In Claude Code, use a hiapi key as HIAPI_API_KEY (tested) or an OPENAI_API_KEY (written, not yet tested).
- A character image, or a written description of the character.
In Claude Code, place the repo folder at ~/.claude/skills/animated-sticker-studio/, or tell Claude to install it with the message below. For Codex, place the folder at ~/.codex/skills/animated-sticker-studio/. The folder already includes agents/openai.yaml.
Tell Claude Code Install this skill: https://github.com/fantuan-lab/animated-sticker-studioThe scripts need Python 3 and Pillow. Run the command below to install Pillow. The Telegram export and the MP4 preview also need ffmpeg.
Install Pillow pip install PillowYou only need one route. Codex needs nothing, because it has a built-in image tool. In Claude Code, the recommended route is a hiapi key set as the variable named below. It was used for the Onigiri Cat example pack (gpt-image-2, 1K transparent, about $0.04 per sheet). If you already have an OpenAI or compatible-gateway key, set OPENAI_API_KEY instead, but that path is not yet tested.
Hiapi key variable name HIAPI_API_KEYSend your agent the character image along with the request below. The README suggests variations: '24 office-worker stickers', 'use my own captions', 'pixel style', or 'also export a Telegram version'. If you have no image, describe the character and the skill draws a reference sheet for you to approve first.
Paste into your agent Use animated-sticker-studio to make 16 animated stickers of this character for the WeChat Sticker Platform.To see the output without spending image credits, go to the example folder in the repo and run the build script. The cut sprites are already in the repo, so nothing is generated.
Run from the repo rootExample assembled from the docs cd examples/onigiri-cat python3 ../../scripts/build_pack.py manifest.json out --platform generic,wechat
Watch out for
- The OPENAI_API_KEY route is written but not yet tested. The hiapi route is the tested one.
- Without ffmpeg, the Telegram export and the MP4 preview fail.
- Agents can't watch a GIF, so the skill checks each sticker as 4 frames on alternating light and dark rows. It flags clipping, halos on dark mode, wrong motion and near-duplicate poses. Still check the final look yourself in the offline preview page.
Questions
- Which platforms can it export for?
- Generic GIF, WeChat Sticker Platform, Telegram, LINE, Discord, Slack and WhatsApp. It follows each platform's size, format, frame count, duration and byte limits, and re-encodes files until they fit.
- What if I have no character image?
- Describe the character instead. The skill draws a reference sheet, and you approve it before the pack is made.
- How many image generations does a pack need?
- The 24-sticker Onigiri Cat example used 4 image-generation calls: 1 character sheet and 3 pose sheets.
- Does it only slide and squash a still image?
- No. Gesture stickers get 2 to 4 drawn keyframes, such as wave, clap, type, eat, cry and yawn, layered with procedural motion.