STLGenie

JPG to STL — Convert a Photo into a 3D Model for Printing

Who this is for: For anyone photographing something small and wanting it as a 3D model: a figurine idea from a product photo, a pet portrait for a shelf piece, an heirloom you want to be able to print a copy of.

Upload a JPG photo and get an STL model of its subject — AI builds the mesh, the page previews it in 3D, and the download is slicer-ready. 2 free models a day.

Quick answer: JPG to STL converts a photo into a printable 3D model: AI infers the object's shape from the picture, generates a triangle mesh, and gives you STL (for slicers) plus GLB (with color) to download. First model each day is free, no account needed.

Image to 3D model

2/day free
Choose an imagePNG / JPG / JPEG / WebP — one clear subject works best

AI reconstruction · STL + GLB downloads · the image is prepared locally in your browser

Core facts
InputOne JPG photo up to 10MB — one clear subject, even lighting
OutputSTL + GLB, previewed in interactive 3D before download
Best subjectsSingle objects or pets; the camera-facing side is most faithful
Compression noteHigh-quality JPG exports reconstruct cleaner than heavily compressed ones
Free tier2 free models a day, no account, no watermark
PrivacyPhoto never stored — result link expires in 24 hours

How JPG to STL works

JPG is the world's most common photo format, and photos are exactly what the reconstruction model expects: a real object captured with real lighting. The model infers depth from that shading and perspective, builds a triangle mesh, and hands it back for a 3D preview, with STL and GLB downloads.

JPG comes with one caveat worth knowing: the compression. Strong JPG compression leaves blocky artifacts in flat areas and around edges, and the reconstruction model inherits some of that noise as surface texture. Export at high quality (or use PNG when available) and the mesh comes out cleaner. Most phone photos at default settings are fine — modern encoders are gentle.

For the best results with a photo: fill the frame with one object, avoid harsh shadows, and include a bit of the object's sides — a slight three-quarter angle gives the model more information than a dead-on shot. The back that the camera never saw will be a plausible reconstruction, not a measurement — good for display pieces, not for mechanical parts that mate with something.

Common uses

  • A photographed toy or figurine becoming a spare copy
  • A pet photo becoming a desk-size sculpture
  • A vase or mug photo becoming a shape reference
  • An old object photographed for a print-in-place replica study
  • A product photo becoming a miniature for a pitch display

What single-photo reconstruction can and cannot do

Can: recognizable shape, display-quality proportions, textured GLB output. Cannot: measured accuracy, hidden geometry, or mechanical tolerances — one photo contains one viewpoint, and the model invents the rest plausibly. If you need dimensional accuracy, photogrammetry from 20+ photos is the right tool; if you need a display piece that looks like the picture, this is it.

Frequently Asked Questions

How does converting an image to STL actually work?
An AI 3D reconstruction model looks at your image the way a sculptor reads a photo: it infers depth and shape from shading, perspective and familiar objects, then generates a triangle mesh matching what it inferred. The mesh is delivered as a 3D file you preview here and download — as STL for printing, or GLB to keep the color.
What makes a good input image?
A single object, clearly lit, filling most of the frame, shot from a slight angle rather than dead-on. Simplify backgrounds help the model separate object from noise. Photos of real objects under even lighting are what the model is trained on. Photos of people work for figurines; flat artwork works for coasters and stamps; product shots work for miniature replicas.
Is the result printable right away?
Usually it needs a slicer health check like any generated mesh: open it in Bambu Studio / PrusaSlicer / Cura, run the repairs if any are flagged, and check wall thickness and scale. The STL you download is a real triangle mesh, not a placeholder — but there is no such thing as an AI mesh guaranteed watertight, so 30 seconds in your slicer's repair dialog is the honest advice.
Can I keep the colors?
Not in STL — the format stores pure triangle geometry, by design. That is why every result also offers a GLB download: GLB keeps texture and color, and it is what you want for painting, rendering or web viewing. For most 3D-printing workflows, color comes from the filament you load, not the file.
What does it cost?
Every day includes 2 free full-quality models (no account, no watermark). After that, credit packs: 3 models for $1.99, 10 models for $4.99, 30 models for $12.99. These fund the GPU time the reconstruction model needs — that part cannot run in your browser, unlike the format converters elsewhere on this site, which are free and unlimited.
Where is my image stored?
Nowhere. Your image is resized locally in the browser and passed through our server to the reconstruction engine only to produce your model. We keep no copy — and the result link expires after 24 hours, so download it when it is ready.

More guides

Best Image to STL Converters

An honest look at the image-to-STL tool landscape in 2026: what free tiers hold back, subscription versus one-time pricing, and what to check before choosing.

More related tools

PNG to STL

Drop a PNG, get a printable 3D model — AI reconstruction turns the picture into a triangle mesh you preview in 3D and download as STL or GLB. 2 free models a day.

JPEG to STL

JPEG images to printable STL models — same pipeline as JPG, with a note on why export quality matters for reconstruction. 2 free models a day.

Lithophane Maker

Turn a photo into a lithophane: a 3D-printed panel that reveals the image when backlit. AI builds the relief, you download the STL and print it thin. 2 free models a day.