# Multi-Image to 3D

{% hint style="info" %}
🔒 Multi-Image to 3D is a **Pro feature.**
{% endhint %}

### 🧠 What is Multi-Image to 3D?

Multi-Image to 3D (Multi-View) generates a 3D model by combining **four matching views** of the same object:

* Front
* Left
* Back
* Right

This reduces guessing in side/back areas and improves overall shape consistency.

***

### 🚀 How It Works

{% stepper %}
{% step %}

#### Prepare views

* Prepare 4 views of the same object: Front, Left, Back, Right.
  {% endstep %}

{% step %}

#### Upload images

* Upload each image into the matching slot (Drag & Drop or **Upload**).
  {% endstep %}

{% step %}

#### Choose a preset

* Choose a preset under **Optimize for**:
  * **High Quality**
  * **Low Poly**
  * **3D Print**
    {% endstep %}

{% step %}

#### Generate

* Click **Generate**.
  {% endstep %}

{% step %}

#### Review & export

* Review the result in the viewer (textured/solid/wireframe) and export when ready.

{% hint style="warning" %}
Only .glb, .usdz, and .blend stores textures in the object.
{% endhint %}
{% endstep %}
{% endstepper %}

***

### ✅ Best use cases

* Props and products where side/back shape matters (tools, furniture, devices)
* Reference-based modeling from orthographic sheets
* Fast 3D printing prototypes using the **3D Print** preset
* Objects with unclear depth in a single image

***

### 🔍Tips for best results

#### Keep views consistent (most important)

* Use true **orthographic** renders if possible.
* If using photos, keep the same:
  * camera distance
  * focal length
  * framing
  * object position and scale

{% hint style="info" %}
Consistent views are the single biggest factor for success.
{% endhint %}

#### Use simple backgrounds and lighting

* Plain background (no clutter)
* Even lighting with minimal shadows
* Avoid strong reflections and highlights

#### Keep the object unchanged

* One object only
* No pose/state changes between views (open vs closed, rotated parts, etc.)
* Don’t crop the object

***

### ⚠️ Limitations to Know

* Mismatched angles between views (can cause twisting/bulging/missing parts)
* Reflective or transparent materials (glass, chrome)
* Very thin details (wires, chains)
* Facial features and characters may look rough or “off”
* Generated geometry is **not parametric** — models can’t be edited with sliders like in the Template Editor
* Complex geometry may need cleanup if used in games.
* Snapshots, palettes, and toon shaders are **only available in Template Editor**

***

### 📦 Export formats

Supported export formats:

* `.glb`, `.obj`, `.ply`, `.stl`, `.fbx`, `.usd`, `.usdz`, `.blend`

Recommended formats:

* ✅ **GLB** (best default)
* 🧩 **OBJ** (broad compatibility)
* 🖨️ **STL** (3D printing)

***

### 🔁 Related tools

Use Multi-Image to 3D alongside other Sloyd tools:

* **Text to 3D**: best for fast ideation without references
* **Image to 3D**: best when you only have one image
* **Text to Image:** best for fast concepting and reference generation
* **Template Editor**: best for clean, editable, game-ready assets

***

### 🌍 Who’s It For?

Multi-Image to 3D is best for anyone who needs the shape to be closer to the real object than a single image can provide:

* **Game devs**: better silhouettes and fewer “made-up” side/back details for props
* **3D artists**: a faster starting point when you already have reference views
* **3D printing hobbyists**: get higher quality geometry using the **3D Print** preset
* **Concept artists**: turn orthographic sheets into a usable 3D blockout
* **Students/Educators**: Create 3D without touching complex software

***

### 🧪 Status

{% hint style="warning" %}
Multi-Image to 3D is under active development. Quality will continue to improve over time, and feedback helps guide what to prioritize next.
{% endhint %}

***

### 🔗 Try It Now

Head to [app.sloyd.ai](https://app.sloyd.ai/) → Choose **Multi-Image to 3D** in the **Create** section.


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://sloyd.gitbook.io/documentation/creating-3d/multi-image-to-3d.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.