The “turn yourself into an action figure” trend is fun on screen, but it’s even better when you can hold it. Here’s the full workflow I use to go from a photo to a 3D printed figurine: ChatGPT for the character, a free image-to-3D tool for the model, and a few slicer tricks for a clean finish.
The short version
- Photo: a clear waist-up or head shot in good light.
- Character: ChatGPT with the prompt below, always on a plain white background.
- 3D model: MakerWorld’s MakerLab image-to-3D gave me the best results; generating is free, exporting costs a few credits.
- Print: add a small base, scale to about 125 mm tall, tree supports, variable layer height.
Step 1: Create the character in ChatGPT
Upload your photo to a new chat with this prompt, replacing XXXX with a short description (for example “the man”):
Transform this photo of the XXXX into a stylized 3D cartoon figure with a soft appearance. The character should have large, expressive eyes, a small nose, chubby cheeks, and smooth, plastic-like skin, similar to a high-quality CGI animation or collectible figurine. Maintain a friendly and cheerful expression. Use simple, neat hair that matches the original style or suits the character’s look. Dress the figure in clothing similar to the picture. The pose should be natural and endearing, standing with hands in pockets or crossed arms. Use a plain white background with soft studio lighting.
- Keep the white background. It helps the 3D tool separate the figure from the background.
- If the figure gets cropped, ask it to “make sure the character is not cut off and fits in the frame”.
- Change the style by editing the prompt: a superhero costume, an angry expression, or “a low poly model style character” for a faceted sculpture look.
- If a conversation goes off the rails after a few edits, start a new chat rather than fighting it.
Step 2: Turn the image into a 3D model
I tested Meshy, PromeAI and MakerWorld. Meshy was decent but odd with colours; PromeAI struggled with the back of models. MakerWorld’s image-to-3D (MakerLab tab, Experiments section) was the clear winner: clean backs, no shading glitches, even sensible detail like paws under a cat’s feet.
- Create a MakerWorld account (setting up your profile earns points you can swap for MakerLab credits).
- Generate new model, drop in your image, wait 3 to 4 minutes.
- Export as OBJ. Generating was free and exporting cost 2 credits during the beta.
Want more control over detail? My newer guide covers turning any image into a 3D print with Hi3D.
Step 3: Add a base (optional)
Figures with rounded feet tend to topple. In Blender (free) I import the OBJ, stand it upright, add a squashed cylinder under the feet, sink it slightly into the model, and export both together as STL with “selection only” ticked. Keep the base compact: my first one was too wide and forced the tree supports out at awkward angles. Alternatively, slice the feet flat and skip the base.
Step 4: Slice and print
- Scale to about 150% (roughly 125 mm tall).
- Supports: tree (auto), slim, build plate only. Keep support contact away from the face.
- Variable layer height: adaptive (quality side), then smooth. This is the single biggest improvement to the surface finish.
- Mine took 2.5 to 3.5 hours and about 50 g each. Silk PLA left the cleanest finish; marble PLA showed a little artefacting.
Want a figure printed without the setup? I print custom figurines and gifts on the Gold Coast, shipped Australia-wide.