# Image to 3D

> Turn a photo, sketch or render into a 3D model at /design — accepted formats, single image versus multi-view, drawing a picture first, what photographs well, and what it costs.

Source: https://x3dstudios.com/docs/studio/image-to-3d  ·  Updated: 2026-10-17

Image to 3D reconstructs a model from a picture instead of a description. It is the same studio at /design — switch the input tab from Text to Image, Image + Text or Multi-view, add your files, and generate. Everything after that is identical to a text prompt: the mesh is repaired, measured, normalised to a printable size and quoted.

Use it when you already know what the thing looks like: a product photo, a concept render, a hand-drawn sketch, a figurine on a table. It is the better path whenever describing the shape in words would take longer than showing it.

## Accepted images

|  | Limit |
| --- | --- |
| Formats | PNG, JPEG, WEBP |
| Size | 10MB per image |
| Single image | Exactly one file |
| Multi-view | 2 to 4 files |

The uploader filters by type in the browser, and the server checks type and size again before a credit is spent. A wrong type is rejected with "unsupported image type", an oversized single image with "image exceeds 10MB limit" (a multi-view set says "an image exceeds the 10MB limit"), and a multi-view set outside the range with "multiview needs 2 to 4 images". A single image is then read by its contents rather than its name: a HEIC, AVIF, BMP or SVG file labelled as something else is refused with advice on what to export instead and "You were not charged.", and a TIFF is converted for you. None of those cost you anything.

> **Resolution is not the limiting factor** — A sharp 1500px photo of a well-lit object beats a 6000px photo of a cluttered shelf. Spend the effort on the subject, not the sensor.

## Three ways in

| Mode | Takes | Prompt | Use it for |
| --- | --- | --- | --- |
| Image | One image | None | Reconstructing what is in the picture, as it is |
| Image + Text | One image and a prompt | Required | Steering the result — "make it hexagonal and taller" |
| Multi-view | 2 to 4 images | None | Objects with a back and sides you care about |

The three differ in more than the file count. In Image mode there is no prompt box and no words reach the generator: the reconstruction is made from the picture alone, and the generation is named in your history after the subject a vision check sees in the photo, or the filename if it cannot tell. In Image + Text your words are required and go to the generator as you wrote them, up to 400 characters, followed by the subject the vision check named and an instruction to match the reference image exactly. Nothing rewrites them, so they steer the reconstruction rather than replacing it.

Multi-view sends every view together in one request and is pinned to Tripo — it is the only engine here that reconstructs from several images, so engine routing is skipped for it, and a TRELLIS.2 pin is ignored. The generation is recorded as "Multiview from N photos".

> **Multi-view is not photogrammetry** — It does not need overlapping frames, a turntable or a calibrated rig. Two to four clean views of the same object is the whole input. Front and back is worth more than four angles ten degrees apart.

## Taking the photo

1. **One subject, nothing else** — Put the object on a plain surface with nothing behind it. A second object in frame gets reconstructed into the same mesh, and separating them afterwards is more work than reshooting.

2. **Contrast against the background** — A white part on a white table has no silhouette to read. Use a background the object clearly is not — a dark cloth for pale objects, a light sheet of paper for dark ones.

3. **Flat, even light** — Overcast daylight or a diffused lamp. Hard shadows get read as geometry, and a blown-out highlight erases the surface underneath it. Avoid direct flash.

4. **Shoot slightly above, straight on** — A three-quarter view a little above eye level shows the top and one side at once, which gives the reconstruction more to work with than a dead-flat elevation.

5. **Fill the frame, keep it sharp** — Get close enough that the object fills most of the picture, and check focus before you upload. Motion blur turns into soft, smeared geometry.

6. **For multi-view, move around the object** — Front and back first, then the two sides if you have them. Keep the lighting and the distance roughly the same between shots.

## What reconstructs well, and what does not

| Works | Struggles |
| --- | --- |
| Single solid objects with a clear outline | Scenes, or two objects in one frame |
| Figurines, toys, ornaments, characters | Wireframe, mesh and lattice structures |
| Product shots on a plain background | Glass, chrome and mirror finishes |
| Chunky mechanical shapes — housings, brackets, knobs | Anything with a specified fit or thread |
| Concept renders and clean 3D turntables | Photos with heavy shadow or a busy background |
| Line-art sketches of a single object | Hair, fur, fabric folds and other fine detail |
| Symmetrical objects seen from the front | Deeply concave shapes hidden from every camera |

The pattern is simple: the generator can only model what a camera could see. A shape that hides its own interior — a deep bowl photographed from the side, the inside of a bottle, the underside of a base — comes back guessed. Multi-view helps, but it does not see through walls.

## What happens to the photo first

- A vision check looks at the picture first: what the subject is, whether the background is plain or cluttered, and whether the photo is of a flat picture.
- A photo of a print, a screen or a framed image is refused as flat_image when the check is sure of it, and you are not charged. Photograph the object itself, or draw one in the image studio below.
- With Remove background on, the default, a cluttered background or a second subject is cut away and the main object placed on plain white. A photo that is already clean goes through untouched, and if the cut-out fails the original is used, so it can never make a result worse.
- On Auto the image goes to Tripo v3.1. TRELLIS.2 is used only when you pin it, and it takes the picture alone, with no words. Hi-Fi on an image asks Tripo for its detailed geometry and textures.

## No picture? Draw one first

The image studio on the design page makes the picture for you. Describe the object and press Generate image · 1 credit. Change something takes a sentence and Apply · 1 credit redraws the picture with that change; Undo steps back and Start over clears it. When it looks right, Use this for 3D switches the studio to Image + Text with the picture and your description, so pose, silhouette and props are settled before a credit goes on the 3D run. A picture that fails, or comes back empty, gives its credit back.

## Credits and time

| Run | Credits | Typical time |
| --- | --- | --- |
| Image, Fast Draft | 2 | 45–90 seconds |
| Image, Hi-Fi | 3 | 45–90 seconds |
| Image + Text, Fast Draft | 2 | 45–90 seconds |
| Image + Text, Hi-Fi | 3 | 45–90 seconds |
| Multi-view, Fast Draft | 2 | Several minutes — it runs on Tripo |
| Multi-view, Hi-Fi | 3 | Several minutes — it runs on Tripo |

*Add 1 for High mesh quality or 4k textures, and 2 for Ultra, 8k or a refinement pass, which only changes the result on Meshy (see /docs/studio/design). The times are the studio's own estimate; Tripo, which image runs use on Auto, can take several minutes.*

Image runs skip the prompt rewrite and the routing agent whichever mode you pick, so a Fast Draft image is not the shortcut a Fast Draft text prompt is: what draft buys you on the image path is the lower base cost, while Hi-Fi buys Tripo's detailed geometry and textures. Either way an image run reaches an engine that can texture, so it returns PBR maps when textures are on. A text draft cannot.

Credits are charged before the generator is called and refunded automatically if the run errors, including when the model was made but could not be saved, which cancels the run rather than handing you a link that will expire. Generation is limited to 20 runs an hour per account. Full detail is on /docs/studio/credits.

## What you get back

The same result as a text generation: a repaired mesh normalised to half a Bambu H2S plate (170 × 160 × 170mm), a mesh review with the measured geometry behind it — watertightness, volume, open boundary edges, non-manifold edges, triangle count, bounding box, thinnest dimension, steepest overhang — and a print estimate for one PLA unit at standard quality. Downloads are STL, 3MF and OBJ.

A model whose review comes back as needs redesign cannot be sent to checkout. Reconstructions from a single image fail that check more often than text generations do, usually because a shape the camera never saw came back open. When it happens, the cheapest fix is another photo from a different angle and a multi-view run, rather than the same picture again.

> **Check the mesh before you order** — A reconstruction can be watertight, well-oriented and still wrong — a smoothed-over hole, a flattened underside, a detail the light hid. The review scores geometry, not resemblance. Rotate the model in the viewer first.

---

- [Photo to 3D print](https://x3dstudios.com/photo-to-3d-print) — What a photo to print costs, how big it comes out and what the AI cannot promise.
- [AI Design Studio](https://x3dstudios.com/docs/studio/design) — Modes, engines, prompt writing and the full credit table.
- [Printability checks](https://x3dstudios.com/docs/studio/printability) — What each measurement in the mesh review means.
- [Upload and print](https://x3dstudios.com/docs/studio/upload) — Send a finished model to the farm, or upload your own STL, GLB or 3MF.
- [Credits](https://x3dstudios.com/docs/studio/credits) — Balances, refunds, top-ups and subscription allowances.

[Try image to 3D](https://x3dstudios.com/design)

## Frequently asked

**What image formats can I upload for image-to-3D?**

PNG, JPEG and WEBP, up to 10MB per image. Anything else is rejected before a credit is spent, and a single image is also checked by its contents, so an iPhone HEIC photo renamed to .jpg is caught and you are told how to convert it. Multi-view takes 2 to 4 images, each under the same 10MB limit.

**How many photos do I need for a good result?**

One is enough for a shape you mostly see from the front. Two to four views is better for anything with a distinct back or sides — front and back beats two near-identical angles, because the model only stops guessing where a camera actually looked.

**How much does image-to-3D cost in credits?**

An image input adds 1 credit to the base cost, so a Fast Draft image run is 2 credits and Hi-Fi is 3. High or Ultra mesh quality and 4k or 8k textures add more on top. Drawing the picture first in the image studio costs 1 credit per picture.

**Does the background of my photo matter?**

It matters less than it did. With Remove background on, which is the default, a photo whose background is cluttered or holds more than one subject has the main object cut out onto plain white before it is reconstructed. A plain, contrasting surface is still the best start. A photo of a flat picture, a screen or a framed print is refused and costs nothing, because it would come back as a slab with the picture on it.