Sticker Pack Maker

What you're buying: a download package (skill instructions + a Python script) that runs on your own computer inside your own AI agent — it is not a cloud service.
You need: an AI agent that supports skills (e.g. Claude Code or Codex) — its subscription or API costs are yours; Python 3.9+ with pillow and numpy; internet connection.
Third-party services: Pollinations.ai character images (free tier: about 1 image per 45 s; optional paid key). Only English prompt text (character look + pose) is sent — no photos or personal data.

Describe your original character — get a whole sticker set ready to upload. No manual background removal, cropping or measuring.

What you get

• 🧩 Transparent PNG stickers: 01.png … NN.png (8 / 16 / 24 / 32 / 40), each up to 370×320, even width and height, ~10 px margin, under 1 MB
• ⭐ main.png (240×240) and tab.png (96×74)
• 📦 Upload zip with only those files — matches the common static chat-sticker store spec (e.g. LINE Creators Market)
• 🖼️ Preview sheet on a checkerboard so you can judge the cut-out at a glance

Example

Mochi Cat Daily, 8 stickers with Traditional Chinese captions (Hi~, Thanks!, OK!, Boo-hoo, Haha, Good night, Hmph!, Love you): transparent PNGs from 230 to 284 × 320 px (all even, within 370×320), 51–65 KB each, white outline and red captions, plus main.png, tab.png and a 10-file upload zip. Spec check: "8 stickers: PASS". Reverse test — odd sizes, over 320 px tall, non-transparent background, wrong main size, missing tab, 9 stickers — every error was caught as FAIL.

Try it — copy a prompt

• "Design an original round hamster mascot and make 16 stickers with English captions for chatting with friends."
• "Make 8 stickers of my character 'Bao the dumpling' with Japanese captions for coworkers."

Why buy it

One payment, as many sticker sets as you can dream up — no per-set fees and no hours of manual cut-outs.

Who it's for

Sticker creators selling on chat-app sticker stores, illustrators, small brands with a mascot, and anyone who wants custom stickers for friends and family.

Features

• Automatic background removal: detects the background colour from the edges (white or the pastel backgrounds AI often gives), removes only background connected to the border — so white eyes or bellies are kept — and clears ground shadows too
• White outline + captions: sticker-style white border; caption fonts for Traditional Chinese, English and Japanese
• Spec validator: checks count, file names and order, size, even dimensions, transparency, file size and main/tab sizes — any error means it isn't done
• Phrase sets: common lines for friends, coworkers, couples and family (Chinese / English / Japanese)
• Redo only the bad ones: --redo 3,7 regenerates just those stickers; changing captions or colours needs no new images

How it works

1. Describe your original character (or ask the agent to design one), how many stickers, caption language and use case.
2. The agent plans each pose and line for you to approve.
3. The script generates, removes backgrounds, outlines, captions, resizes to spec, validates and zips.
4. The agent checks the preview and redoes any bad stickers before delivery.

Requirements

• Python 3.9+
• pip install -U pillow numpy
• Internet connection

Speed

• Free image tier: about 1 image per 45 s — 8 stickers in about 6 minutes, 40 in about 30 (resumable)
• Faster: get a key at enter.pollinations.ai and set it as the POLLINATIONS_KEY environment variable

Limitations (please read)

• AI-generated content: the character looks similar across a set but not identical (eye colour or markings may vary slightly). On the free tier the model follows pose instructions less reliably, so some stickers may need --redo.
• Static stickers only — no animated, emoji or sound stickers.
• Background removal assumes a plain background; very light characters may need a careful edge check.
• Original characters only — no Pokémon, Sanrio, Disney or anime characters or look-alikes, and no real people.
• It doesn't submit to any store; review standards and AI-content rules vary and change, so check them yourself.
• No guarantee of approval or sales. Not affiliated with LINE or any sticker store.

Files

• SKILL.md — agent workflow
• scripts/make_stickers.py — generation, background removal, outline, captions, spec output, validation, zip
• references/phrase-sets.md — common phrase sets
• references/design-rules.md — sticker design principles
• references/store-rules.md — specs and common rejection reasons
• references/example_zh.json — 8-sticker example