Quick answer
Upload a clear, waist-up photo to ChatGPT or Gemini and paste one of the prompts below. Spell out the name and tagline in quotation marks, list three accessories by name, and ask for the face to match the photo. Start a new chat for each attempt instead of editing the same image again and again; that’s what keeps the likeness. Action Figure AI does the boxed version without a prompt.
Ask a chatbot for “an action figure of me” and you get a toy, but rarely your toy. The good results in the 2025 trend came from prompts that read like a product brief: what the packaging looks like, what’s written on it, which items come in the box, and how the figure is posed. The prompts below follow that pattern. They’re for general tools, not the template on our site.
The boxed figure (ChatGPT)
Turn the person in this photo into a realistic collectible action figure, sealed in retail blister packaging. The figure has glossy plastic skin and visible joints at the shoulders, elbows and knees, and keeps the face, hairstyle and outfit from the photo. The card backing is [colour], with the name "[YOUR NAME]" in bold at the top and the tagline "[TAGLINE]" underneath. Inside the bubble, next to the figure, are three accessories in their own moulded slots: [ITEM 1], [ITEM 2] and [ITEM 3]. Studio product lighting, the pack hanging on a toy shop peg, square image.
Fill the brackets before you send it. Short names spell better than long ones; “DR. MAYA” survives where “Doctor Maya Rodríguez-Hall” gets mangled.
The desk figurine (Gemini)
Make a 1/7 scale painted figurine of the person in this photo, standing on a clear round acrylic base on a tidy wooden desk. Keep their face, hair and clothes. Behind it, a computer monitor shows the figure being sculpted in 3D software, and a printed toy box with artwork of the same character stands beside the base. Soft daylight from a window, shallow depth of field, photorealistic.
This is the look from the September 2025 Gemini wave. Upload the photo in Gemini and paste the prompt. It needs no text on the box, which makes it the easier of the two to get right.
Picking the three accessories
The accessories are the punchline, so name real objects rather than ideas. “A coffee mug, a laptop and a lanyard” draws cleanly; “something that shows I work hard” does not.
Nurse
Instantly readable.
a stethoscope, a water bottle, comfortable clogs
Remote worker
The self-deprecating one.
a laptop, slippers, a giant iced coffee
Runner
For the race-day post.
a medal, an energy gel, a pair of running shoes
Keeping it looking like you
- Upload one sharp, well-lit photo, face straight on. Group shots and selfies at arm’s length both lose detail.
- Write “keep the face exactly as in the photo” in the prompt; it helps more than adjectives about your looks.
- If the first result is off, start a fresh chat with the same photo and prompt. Asking for fix after fix on one image is how faces drift into someone else.
- Check the box text letter by letter before you post. If one word is wrong, fixing it in an editor is quicker than re-rolling.
Free tiers of both tools cap how many images you can make in a day, and the caps change. If you hit one mid-attempt, the next day’s first try is usually better anyway, with a cleaner chat.
No prompt needed
On Action Figure AI the blister pack, shelf and plastic finish are fixed, the accessories are picked from your outfit, and the card is left blank. It’s 4 credits for two images, and the first photo trend on a new account is free.
- 1Open Action Figure AI and add a waist-up photo in a telling outfit.
- 2Check the price on the button, then press make.
- 3Download both images from Assets and add your text in the caption.
Common problems
The name on the box is misspelled
Put the exact text in quotation marks and keep it to one or two words. If it still slips, correct it in a photo editor.
It looks like a generic doll, not me
Add “keep the face exactly as in the photo”, use a sharper photo, and start a new chat rather than editing the old image.
The accessories float outside the box
Say “inside the bubble, in their own moulded slots”. Without it, tools scatter them around the pack.
FAQ
What is the ChatGPT action figure prompt?
Describe the figure, the blister pack, the text on the card and three accessories by name, and ask it to keep your face. The full prompt is above.
Does the action figure trend work in Gemini?
Yes. Gemini is known for the desk-figurine version on a clear stand, but the boxed prompt works there too.
Is the action figure trend free in ChatGPT?
Free accounts can make images, within a daily limit that changes over time. A few attempts usually fit in one day.
Is this the prompt Reelfie uses?
No. The template has its own setup and keeps the card blank. These prompts are for other tools.
Make your first one
Action Figure AI: You, boxed as a collectible.
Try Action Figure AIYour photos are deleted 7 days after upload and never used to train AI.
