Articles on: Redraw Features

3D Object

3D Objects: turn images and descriptions into three-dimensional models in Redraw

Redraw's 3D Object Generator turns images and descriptions into interactive three-dimensional models, making it easier to visualize shapes, volumes and proportions within your project.
With this feature, it becomes simpler to explore different angles of the same object, enrich presentations and produce more complete and dynamic visual materials.
The feature is available in Redraw's side menu, under 3D Objects.


How the 3D Objects feature works

Redraw's 3D Object Generator offers different artificial intelligence models for creating three-dimensional models.
Besides Redraw's own model, you can also choose between Meshy v6, Rodin V2.5 and Tripo H3.1, technologies recognized in the market that expand your generation possibilities.
The Redraw model remains available and has received a recent update, with a new generation model, detail levels and improvements to the experience while you wait.


What changed in the Redraw model

The platform's own model has been updated and now offers:

  • a new generation model, with higher quality objects;
  • three detail levels: Light, Balanced and Detailed;
  • a new point cloud animation during processing;
  • generation running in the background, so you can keep using Redraw;
  • the final object available to rotate, view and download.

The other models remain available as usual, with their own settings.


Available creation modes

Depending on the model you choose, you can create three-dimensional objects using different input formats.


Image to 3D object

Upload a single image so the artificial intelligence uses that reference to create the model.
This mode is recommended when you already have a photograph, render or illustration of the object.

Text to 3D object

Describe the object using a text prompt.
The more detailed the description, the greater the chances of getting a result close to your idea.
In the Redraw model, before the three-dimensional generation the system displays a preview image so you can confirm the direction is correct. Once approved, the 3D object is generated.

Multi image

Upload up to four images of the same object, showing different angles.
The first image you upload is the main reference for the generation, so it should be as complete and sharp as possible. The following images work as complementary material.
All images must represent exactly the same object. Differences in shape, proportion, color or finish make interpretation harder.


Available models

The 3D Object Generator currently has four artificial intelligence models.

Redraw

Redraw's own model remains available and offers the modes:

  • Image
  • Text
  • Multi image

After the update, it also lets you choose between the Light, Balanced and Detailed levels, displays the point cloud during processing and runs the generation in the background.
It's a versatile option for different types of generation within the platform.

Meshy v6

Meshy v6 is focused on generations based on visual references.
Available modes:

  • Image
  • Multi image

It's recommended when you have one or more images of the object and want to use them as the main reference.

Rodin V2.5

Rodin V2.5 lets you create objects from image or text.
Available modes:

  • Image
  • Text

It also offers quality options for different needs:

  • Fast
  • Normal

In addition, you can enable the High definition option, which prioritizes a result with a greater level of detail.
It's a good choice for balancing speed and quality during generation.

Tripo H3.1

Tripo H3.1 brings together different creation formats and advanced detailing resources.
Available modes:

  • Image
  • Text
  • Multi image

It also offers the options:

  • HD Texture
  • Detailed Geometry

The HD Texture option improves the appearance of materials, colors and finishes.
Detailed Geometry prioritizes curves, reliefs and small structural details of the model.


Detail levels in the Redraw model

When generating with the Redraw model, you choose between three levels, which define the balance between visual richness and file weight.

Light

  • Smaller file size.
  • Opens and rotates faster.
  • Fewer details on the object.

Recommended for scenes with many objects or when you need a file that's quicker to handle.

Balanced

  • Good level of detail.
  • File without excessive weight.

This is the default option and covers most cases.

Detailed

  • Higher level of detail.
  • Larger file, which may take longer to open and move.

Recommended for close-ups and objects that need more visual richness.
The level selector applies to generations made with the Redraw model. In the other models, quality and detailing options follow each technology's own settings, such as Fast, Normal and High definition in Rodin V2.5 and HD Texture and Detailed Geometry in Tripo H3.1.


The point cloud during generation

In generations made with the Redraw model, the system displays an animation of points formed from the outline of the image you uploaded. It indicates that the generation is in progress.
The point cloud is based on a single image and is not a preview of the final object. It may look different from the result and should not be used to judge the object's quality before the generation finishes.


Follow the generation while you use Redraw

The generation happens in the background and you don't need to stay on the screen. While the object is being processed, you can browse Redraw, start another generation and use the other features. You'll get a notification when the object is ready.
Generation usually takes between 1.5 and 4 minutes.


How to choose the right model

Each model has its own characteristics.

  • Choose Redraw when you want to use the platform's default model to generate objects from image, text or multiple images, with control over the detail level.
  • Choose Meshy v6 when you have visual references and want to take advantage of different angles of the object.
  • Choose Rodin V2.5 when you want to generate from image or text and control quality between Fast, Normal and High definition.
  • Choose Tripo H3.1 when you need to combine different input formats and want more control over textures and geometry details.

There is no single ideal model for every project. Each technology interprets references differently and may produce distinct results.


How to get better results

For image generation

To make interpretation easier for the artificial intelligence:

  • Use sharp images.
  • Keep only one main object in the image.
  • Center the object.
  • Avoid cutting off important parts.
  • Use good lighting.
  • Prefer simple backgrounds.
  • Avoid elements covering the object.

For multi image generation

Start with the most complete and informative image of the object, since it is the main reference for the generation.
Then use images showing:

  • Front;
  • Sides;
  • Back;
  • Three-quarter perspective.

All images must represent exactly the same object.

For text generation

Describe the object in as much detail as possible.
Include information such as:

  • Object type;
  • Shape;
  • Material;
  • Color;
  • Finish;
  • Style;
  • Proportions;
  • Important details.

Example:

A light wood chair with a curved backrest, a gray fabric upholstered seat and a Scandinavian style.


What to expect from the result

The 3D object is created from the interpretation of the image or description. The model needs to interpret the object's volume and reconstruct parts that don't appear in the reference, such as the back and hidden details.
Because of that, it's expected that the result shows differences compared to the original image, and that two generations made from the same reference produce distinct variations.
Generated objects are delivered as a single piece, without selection or separation of specific parts.


What you can do with 3D Objects

With the 3D Objects feature you can:

  • Explore different angles of the same object.
  • Better understand volumes and proportions.
  • Create objects from images or descriptions.
  • Use different AI models depending on the project.
  • Enrich presentations.
  • Produce more complete materials for analysis and visualization.


Download format

After generation, the object can be downloaded in GLB format, compatible with several 3D software packages and viewers.


Frequently asked questions

Will the Redraw model be removed?

No. The Redraw model remains available as usual and has been updated with a new generation model.

Do the new models replace Redraw?

No. They expand the options in the 3D Object Generator and can be chosen according to each project's needs.

What changed in the Redraw model?

The update brought higher generation quality, three detail levels, the point cloud during processing and background generation.

What's the difference between the Light, Balanced and Detailed levels?

Light produces a lighter file that's faster to handle, Balanced covers most cases, and Detailed prioritizes the amount of detail, resulting in a larger file.

What are the dots shown during generation?

That's the point cloud, an animation based on the uploaded image that appears while the object is being processed. It isn't a faithful preview of the final object.

Do I need to stay on the screen while the object is generated?

No. The generation happens in the background and you can keep using Redraw normally.

How long does a generation take?

Usually between 1.5 and 4 minutes.

I generated twice with the same image and the result changed. Is that normal?

Yes. Each generation may produce its own variation, even when the reference is the same.

Can I separate parts of the generated object?

No. Objects are delivered as a single piece.

Which model accepts multiple images?

Redraw, Meshy v6 and Tripo H3.1 offer the Multi image mode.

Which model allows generating from text?

Redraw, Rodin V2.5 and Tripo H3.1 let you create objects from text descriptions.

Which model has fast generation?

Rodin V2.5 offers the Fast and Normal options.

What does the High definition option do?

It prioritizes a generation with a greater level of detail.

What is HD Texture?

It's a Tripo H3.1 option that improves the visual quality of surfaces, materials and finishes.

What is Detailed Geometry?

It's a Tripo H3.1 option that increases the level of detail in the object's three-dimensional structure.

Can I turn an image into a 3D object?

Yes. All models available in Redraw let you generate three-dimensional objects from an image, as long as they offer the Image mode.

Can I try more than one model?

Yes. Since each technology interprets references differently, trying different models can help you find the most suitable result.


Conclusion

Redraw's 3D Objects feature lets you turn images and descriptions into three-dimensional models using different artificial intelligence technologies.
Besides Redraw's own model, now updated with higher quality and detail levels, you can also choose between Meshy v6, Rodin V2.5 and Tripo H3.1, selecting the most suitable option for each project.
With different creation modes, quality settings and detailing resources, the 3D Object Generator offers more flexibility to turn ideas into three-dimensional models in a practical and efficient way.

Updated on: 21/09/2026