> For the complete documentation index, see [llms.txt](https://sloyd.gitbook.io/documentation/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://sloyd.gitbook.io/documentation/creating-3d/image-to-3d/the-easiest-way-to-turn-your-images-into-3d-models.md).

# The easiest way to turn your images into 3D models

{% stepper %}
{% step %}

### Step 1: Open Image to 3D

* Go to Sloyd’s **Image to 3D** page.
* Drag and drop your image into the upload box.
  {% endstep %}

{% step %}

### Step 2: Choose your settings

* Pick the settings that match your use case.
  {% endstep %}

{% step %}

### Step 3: P**review the result** 

* Use the **in-browser preview** to check the mesh.
  {% endstep %}

{% step %}

### Step 4: Export the right format

Choose the export format based on where you’ll use the model:

* **GLB or FBX**: recommended for game development
* **STL**: recommended for 3D printing

Download the file in the format that fits your pipeline.
{% endstep %}
{% endstepper %}

{% hint style="info" %}

### Pro tips for best results

#### Use the right kind of image

This works best with:

* **High-contrast images**
* **Transparent backgrounds**
* Single, clear subjects like **characters, buildings, or objects**

#### Avoid busy photos

If the image has a lot going on, the AI may misread shapes and produce a messy result.

#### Remove backgrounds before uploading

For best results, remove the background first. You can use any background-removal tool, including:

* ChatGPT (image editing)
* Sloyd’s built-in image edit feature
* Any external tools
  {% endhint %}

***

### Use the model in your 3D software

After exporting, import the file into your tool of choice, for example:

* Blender
* Unreal Engine
* Unity
* Roblox

The result is a clean mesh that follows the shape of your image, ready to use as an asset.

***

### Video tutorial

* [The easiest way to turn your images into 3D models!](https://youtu.be/6dTvG5Abu2w?si=59tFSQsDuQ_LROr8)