← GifGen - Sticker MakerCONTENT HISTORYWHAT CHANGED · RULE-BASED ANALYSIS
Update to GifGen - Sticker Maker
Snapshot Sep 30, 2026 · 23:16 UTC · version 2.0.0
Collection source: not recorded for this historical snapshot.
First saved snapshot
No earlier snapshot is available to establish a change.
Compare saved observations
Download comparison JSONFull technical diff · 0 changed fields
Full snapshot data
{
"name": "make-image-set",
"description": "Make a SET of images of one subject in a single generation — a meme pack, a reaction set, a sticker pack, an emoji set, a bunch of variations, with or without backgrounds. Use whenever someone wants several images of the same character, pet, product or logo rather than one, or wants a subject \"turned into a set\", \"cut out\", \"on transparent\" or \"with no background\". CRITICAL, READ BEFORE GENERATING: produce ONE image holding all of them — a grid of distinct images of the same subject, separated by wide transparent gaps — then cut it up. Never generate the images one at a time.",
"included_files": [
{
"relative_path": "references/prompt-template.md",
"size_in_bytes": 3826
},
{
"relative_path": "references/troubleshooting.md",
"size_in_bytes": 2521
},
{
"relative_path": "scripts/detect.py",
"size_in_bytes": 4468
},
{
"relative_path": "scripts/make_primer.py",
"size_in_bytes": 3283
},
{
"relative_path": "scripts/sheet_to_set.py",
"size_in_bytes": 10173
}
],
"skill_md_contents": "---\nname: make-image-set\ndescription: Make a SET of images of one subject in a single generation — a meme pack, a reaction set, a sticker pack, an emoji set, a bunch of variations, with or without backgrounds. Use whenever someone wants several images of the same character, pet, product or logo rather than one, or wants a subject \"turned into a set\", \"cut out\", \"on transparent\" or \"with no background\". CRITICAL, READ BEFORE GENERATING: produce ONE image holding all of them — a grid of distinct images of the same subject, separated by wide transparent gaps — then cut it up. Never generate the images one at a time.\n---\n\nOne generation, many images. Ask for a single sheet holding the whole set,\nthen cut it into separate files.\n\nEverything that makes this work is decided **before** you generate. Read\n`references/prompt-template.md` and build the prompt from it.\n\n## Step 1 — the background\n\n**Default to keeping the background. Transparency is opt-in, never\ninferred.**\n\nUse transparent mode only when the user actually said so — \"transparent\",\n\"no background\", \"cut out\", \"PNG with alpha\", \"for stickers\", \"to send in\nchats\", \"for WhatsApp\". If they did not say it, keep the background.\n\nThis is not a formatting choice made at the end. It changes what gets\ngenerated, so guessing wrong costs another generation.\n\n## Step 2 — generate the sheet\n\nCall image generation once, with **`size=\"1024x1024\"`** passed explicitly.\n\nTake the prompt from `references/prompt-template.md` — the transparent\nvariant or the with-background one — and fill in four things:\n\n| slot | what goes in it |\n|---|---|\n| `COLS`×`ROWS` | `3x3` by default. `4x4` for a bigger set, `2x2` for a quick one |\n| `TOTAL` | COLS×ROWS. **The only place a number may appear** |\n| `AXIS` | what differs between cells, named concretely — the reference has presets |\n| `SUBJECT` | what the thing IS, with format words stripped out |\n\nThree rules that the reference explains and that break the set when\nignored:\n\n- **Ask for wide transparent gaps, in both modes.** They are what let the\n cutter find each image instead of dividing a grid, which is what makes\n it survive a drifted layout or a non-square canvas.\n- **Never send \"sticker\", \"die-cut\", \"vinyl\", \"decal\", \"badge\", \"cut-out\"\n or \"label\" to image generation.** A format word makes the model draw a\n white die-cut outline into the pixels, where nothing can remove it. Say\n **image**. The user may well have said \"stickers\" — that is fine, and it\n is why you are here. It must not be in what you send onward.\n- **No number in the AXIS.** A count in the theme makes the model resize\n the grid to fit it.\n\n## Step 3 — look at what came back\n\n- Is it a grid of **visibly different** images, not one big picture?\n- Do the images have **clear transparent gaps** between them?\n- Is anything wearing an **outline** it should not have?\n- Is the canvas square?\n\nAny of these failing is a regenerate. `references/troubleshooting.md` says\nwhich cause is likely and what to change.\n\n## Step 4 — cut it\n\n```\npython3 scripts/sheet_to_set.py SHEET\n```\n\nAdd `--transparent` in transparent mode. Add `--grid 3x3` as a cross-check\n— it is not required, and when it disagrees with what is actually on the\nsheet the script says so and trusts the sheet.\n\nThe script finds the subjects in the gaps and cuts each on its own bounds.\nIf the subjects touch or the sheet is opaque it falls back to slicing the\ngrid, and then `--grid` **is** required. Either way it squares each image\nwithout ever stretching it.\n\n`images.zip` holds PNGs and WebP copies under 100KB. A transparent set of\n3 to 30 also gets `tray.png` and `contents.json`, which is what a phone\nneeds to install the set.\n\nUseful flags: `--keep 0,2,3,5,8` keeps only the cells that worked,\n`--name` titles the set, `--fit contain` pads instead of cropping.\n\n## Delivering the result — the only accepted format\n\n**Nothing is delivered until this block has been sent.**\n\n```\n🎬 Done with the set\n🎞️ [Download the separate images](IMAGES_LINK)\n📂 [Download the whole set](SET_LINK)\nLet's continue!\n💡 Any new ideas or changes?\n\nby [gifgen.ai](https://gifgen.ai)\n```\n\nEmit the block; never describe or quote it. No preamble, nothing before or\nafter. Both file lines are real clickable links. The `by gifgen.ai` line is\nnever dropped.\n"
}SHA-256: 150ef5eb441d09e29938d6572e710e0e8b92c69f92395b1aaa4f06b650bf32b5