AI Action Figure and Chibi Prompts: 9 Templates for the Toy Trend

The single most viral AI photo trend right now is people turning themselves into toys. Boxed action figures with accessories, big-headed chibi figures, blister cards with their name printed across the top. It has been running for months, it keeps mutating, and the reason it refuses to die is simple: it is your face, your outfit and your stuff, rendered as a collectible you wish existed.

This is a prompt library for that trend. Nine copy-paste templates covering the boxed figure, the chibi desk figure and the variants people are posting this month, plus the two things that decide whether your result looks like the viral posts or like a melted doll: how you handle the text on the box, and what you feed the generator in the first place.

The six blocks every figure prompt shares

Read fifty viral figure posts and the prompts underneath them all specify the same six things. Miss one and the model fills the gap with something generic:

This is a specialized version of the general prompt structure I broke down in the AI image prompt guide. The same rule applies here as there: concrete beats vague. "Wearing a denim jacket, holding a coffee mug" gives the model something to build; "looking cool" gives it nothing.

All nine templates below work the same way: attach a clear photo of yourself (or your pet), paste the prompt, and swap the bracketed parts. They are written for the current image models in Gemini and ChatGPT, and they degrade gracefully on others.

The boxed action figure

The one that started it all. The packaging is the joke, so give it most of your prompt budget.

Turn the person in this photo into a collectible action figure, posed inside retail blister packaging with a clear plastic bubble on a printed cardboard backing. The card reads "[YOUR NAME]" in bold letters at the top and "[NICKNAME OR ROLE]" in smaller text underneath. The figure wears [THEIR ACTUAL OUTFIT] and stands in a confident pose. Include three small accessories in the bubble beside the figure: [ITEM 1], [ITEM 2], [ITEM 3]. Toy-quality plastic sheen, studio product photography, sharp focus, plain background.

A premium collector edition figure of the person in this photo, displayed in an open window box like a high-end collectible. Matte box with foil accents, the name "[YOUR NAME]" printed on the side panel and "LIMITED EDITION 1 OF 1" on the front corner. The figure is detailed and realistic rather than cartoonish, wearing [OUTFIT], standing on a small display base. Dramatic product lighting against a dark background, shallow depth of field.

A 1990s style action figure of the person in this photo, on a faded retro cardboard blister card with a starburst sticker reading "NEW!" and a price tag of $4.99. The card design uses bold 90s colors and reads "[YOUR NAME]" in a chunky retro font. Slightly worn packaging edges, the figure has the simple joints and glossy paint of a vintage toy, shot like an eBay listing photo on a plain table.

The chibi and desk figures

The softer branch of the trend: oversized head, tiny body, glossy vinyl. These get shared more by people who find the boxed version too self-promotional.

Turn the person in this photo into a cute chibi vinyl figure with an oversized head and small body, standing on a round white base on a wooden desk next to a laptop and a coffee mug. Glossy smooth vinyl material, soft studio lighting, shallow depth of field with a blurred home office background. The figure keeps the person's real hairstyle, glasses and [OUTFIT].

A tiny chibi keychain figure of the person in this photo, attached to a silver keyring, held between a thumb and forefinger to show its small size. Soft rubber material with visible seam lines like a real keychain toy, the person's real hairstyle and [OUTFIT] simplified into the chibi style. Bright natural window light, close-up macro shot.

A chibi vinyl figure set of the person in this photo together with their [PET TYPE], both in matching oversized-head chibi style on a single shared display base. The pet sits beside the person's leg looking up. Glossy toy finish, pastel background, soft even lighting, product photo composition with both figures fully in frame.

Pets, couples and the other twists

Turn the [PET TYPE] in this photo into a collectible action figure in blister packaging. The card reads "[PET NAME]" in bold letters with "[FUNNY TITLE, e.g. PROFESSIONAL NAPPER]" underneath. Include three accessories in the bubble: [FAVORITE TOY], [FOOD BOWL], [ITEM 3]. The figure captures the pet's real fur colors and markings, toy plastic sheen, studio product photo.

A two-pack action figure set of the two people in this photo, posed side by side in shared blister packaging with a cardboard backing that reads "[NAME] & [NAME]" at the top and "THE ORIGINAL DUO" underneath. Each figure wears their real outfit from the photo, and each has one personal accessory beside them: [ITEM FOR PERSON 1] and [ITEM FOR PERSON 2]. Bright toy-aisle product photography.

An action figure playset based on the person in this photo and their job as a [PROFESSION]. The boxed set includes the figure wearing [WORK OUTFIT] plus a themed mini environment: [WORKPLACE PROP 1], [PROP 2], [PROP 3]. The box reads "[NAME]: [PROFESSION] EDITION" with playful toy-brand styling. Bright colors, studio lighting, slight top-down angle showing the whole set.

Getting the box text to come out right

Text is where most figure images fall apart. Image models paint letters rather than typing them, and long phrases come out as alphabet soup. Three rules keep it clean:

• Quote the exact text. Writing: the box reads "MAYA" tells the model precisely which letters to paint. Describing text vaguely ("with her name on it") invites gibberish.
• Keep it short. A name and a two-word subtitle usually survive. A full sentence usually does not. If you want a slogan, expect a couple of retries.
• Fix text with a follow-up, not a redo. In Gemini and ChatGPT you can reply "keep everything the same but fix the text on the box to read MAYA" instead of regenerating from scratch and losing a pose you liked.

Make it postable

Generators mostly hand you a square image around 1024 pixels, which is fine for a feed post but wrong for everything else. The practical chain, all in the browser:

• Going to Stories or Reels? Crop to 9:16 with the crop tool before the platform does its own center-crop and eats your box text.
• Making it a profile picture? The profile picture cropper shows the circle crop before your platform surprises you; figures with box text usually need the figure, not the box, centered.
• PNG downloads from generators run heavy. A pass through the image compressor with a target size gets a multi-megabyte PNG under any platform cap; my viral photo trends post has the exact sizes that matter per platform.
• Want it to look less obviously AI? The film grain treatment works surprisingly well on figure images, especially the retro 90s card, which practically begs for it.

The privacy part nobody reads

Two honest notes before you upload your face to a generator. First, you are uploading your face: check what the service retains and opt out of training where the setting exists. One clear selfie does the job; there is no reason to hand over a whole camera roll. Second, generated images do not carry your location the way phone photos do, but many generators embed invisible provenance marks like SynthID or C2PA credentials in the file. That is arguably a feature, not a bug, but if you want to know what is inside an image, the metadata viewer shows you, and I covered how provenance marks work in how to tell if an image is AI-generated.

Now go make yourself into a toy. The two-pack with a pet is the sleeper hit of the nine; nobody expects the dog to have accessories.