How to Make an AI Figurine From a Photo
The figurine look went mainstream in 2025 and most attempts still fail the same way: the result is a stylised drawing of a person rather than a photograph of a manufactured object. The difference is not the model you use — it is whether the prompt describes an object at all. Here is what actually carries the effect.
Try a figurine template freeName a scale, or nothing else works
A scale like 1/7 or 1/8 is the single most load-bearing phrase in a figurine prompt. It tells the model it is rendering something with a specific physical size, which brings along everything that comes from photographing a small object up close: shallow depth of field, a visible parting line where the mould halves met, and the way light wraps a 25cm figure differently from a person. Without a scale the model has no reason to treat the subject as manufactured, so it defaults to illustrating a human.
Say what it is made of
Injection-moulded PVC, cast resin and polystone all behave differently under light — PVC has a slight surface give and a soft specular roll, resin reads harder and more brittle, polystone reads like painted stone. Naming one is what produces material realism instead of a generic plastic shine. Painted detail matters too: airbrushed shading and hand-painted eye decals are the specifics that read as a real product rather than a 3D render.
The base and packaging sell the context
A clear acrylic base with a printed nameplate, or a blister pack behind a window box, is what tells the viewer this is a collectible rather than a small person standing on a table. It is also the cheapest single upgrade to a mediocre figurine prompt — adding a base changes how the whole frame is interpreted.
Light it like a product, not like a portrait
Product photography means a soft key from slightly above, a fill card lifting the shadow side, and a subtle rim to separate the figure from the backdrop — usually a seamless grey or gradient sweep. Portrait lighting on a figurine is the tell that ruins otherwise good renders: it flatters the face and loses the surface behaviour that makes the object look real.
Chibi is a different decision, not a lesser one
Realistic proportion figurines and chibi (nendoroid-style) figures need different prompts. Chibi deliberately breaks proportion — oversized head, short body, simplified features — and reads as gashapon or blind-box product. If you want that look, ask for it explicitly; if you do not, say 'realistic body proportions', because many models drift toward cuteness when they hear 'figure'.
How to use it
- Pick one clear photo: single subject, face and upper body visible, even light, minimal background clutter.
- Decide the style first — realistic 1/7-scale display figure, or chibi with broken proportions. They are different prompts.
- State the scale, then the material (injection-moulded PVC, cast resin, polystone), then the paint finish.
- Add the base or packaging: clear acrylic base with printed nameplate, or a window box with a blister insert.
- Specify product lighting — soft key above, fill on the shadow side, rim light, seamless backdrop — not portrait lighting.
- If you want a turnaround video instead of a still, say the figure rotates on its base and keep every other clause identical.
- Or skip the writing: the figurine templates already carry all of the above, so one upload is enough.
FAQ
How do I make an AI action figure of myself?
Upload one clear photo of yourself and use a figurine template, or write a prompt that names a scale, a material, a base and product lighting. The scale is the part people leave out, and it is the part that makes the result read as an object instead of a drawing.
Why does my figurine prompt produce a cartoon instead?
Almost always a missing scale and material. Words like "figure" or "toy" alone push a model toward illustration; "1/7 scale injection-moulded PVC figure" pushes it toward product photography.
What is the difference between a figurine and a chibi figure?
Figurine style keeps realistic body proportions and sells the effect through material and lighting. Chibi deliberately breaks proportion — big head, small body, simplified features — closer to a blind-box or gashapon toy.
Can I get a video of the figure rotating?
Yes. Ask for the figure rotating slowly on its base and keep the material, base and lighting clauses unchanged. Several of the figurine templates output a short turnaround instead of a still.
What photo works best as a source?
One subject, face and upper body clearly visible, flat or even lighting, and as little background clutter as possible. Group shots and strong backlighting are the two most common causes of a muddy result.
Do I need to write a prompt at all?
No. The figurine templates already contain the scale, material, base and lighting wording, so a single photo upload is enough. The prompt library shows you the exact text if you want to adapt it.
Templates powered by How To Make Ai Figurine From Photo
