Experience NeRF on your smartphone using the 3D capture tool Luma AI!

Many people have heard of “Luma AI” but don’t know what the app actually is, or are curious about what makes it different compared to other 3D scanning apps available on smartphones (WIDAR, Scaniverse, etc.).
This article explains what Luma AI is, how to use it, and how it differs from other scanning apps.



LumaAI completed scene

This scene can be viewed in the 3D viewer of the web version of Luma AI


View in AR (smartphone)

If you are

viewing this article on a smartphone, please select the button below


View in AR (


) If you are viewing this article on

a computer, scan the QR code below with your phone’s camera


QR code for AR scene

What is Luma AI

? A

3D capture tool that can create and output highly accurate 3D models with a single smartphone.
By using an AI-based image generation technology called NeRF, it is unique in that it allows for photorealistic shading and other expressions that have been difficult with previous 3D scanning applications.

Currently (2022/12/08) there are web and app beta versions, with beta versions only available for iOS users to experience.
Also, output is supported in “GLTF”, “OBJ” and “PLY”.

Luma AI Advantages

High reproducibility of reality


shown in the following tweets, Luma AI can reproduce bright and transparent objects that 3D scanning apps usually have trouble with, and can also represent light reflections and reproduce models in dark environments. https://twitter.com/Donye_NFT/status/1580777978812059648?s=20&t=bGj-_mS9VqHR7QrNDI7t4w

https://twitter.com/ko_tavy/status/1582033819875827712?s=20&t=I-APNiu7enYsHzaDkOuzrQ https://twitter.com/Drei89/status/1583656097214124032?s=20&t=1qB9bFU18FnIte5MOGH4WQ https://twitter.com/nikkiccccc/status/1582541321189068800?s=20&t=nous3pYLsoPlRAACpnDZYA https://twitter.com/gravicle/status/1581150143914516480?s=20&t=qjJdD5yeIzUy_U19-1UidQ



creating a 3D model, a guide is displayed on the subject, and by moving the phone accordingly, a beautiful 3D model can be created.
This eliminates the need to think about how many shots to take of the subject from what angle and the need to press the shutter button every time.


Scanning scene guide


You can create 3D models from videos.

With the

web version of Luma AI, you can create 3D models even from simple videos. https://twitter.


Cons of Luma AI

It takes a long time to create a 3D model

It takes longer to

generate a 3D model than other 3D scanning apps, about 20 minutes


Cannot edit the app


other 3D scanning apps, you can edit the 3D model on the app after it is generated.
However, currently (2022/12/08), Luma AI app does not have such a function. If you want to trim the generated model, you need to output the model and edit it on an external app.


After you

apply, you will receive a Discord invitation from the official Luma AI website


email. If you join the Discord, you will receive a TestFlight link to install the beta version.

When you

launch the app, you will be asked to sign in with your AppleID. Please register using any AppleID.



スキャンシーン1 Scanning Scene 1[/caption]

Align the blue crosshairs on the screen with the center of the object and tap the screen


Scanning Scene 2

After moving to the other side of the object, align the blue crosshairs on the screen with the center of the object again and

tap the screen


Scan Scene 3

Adjust the size and orientation of the white wire box by sliding the blue arrow or yellow measure so that the object fits inside the white wire box


Scan Scene 4

Slowly move around the object to complete the purple loop (line) and also tilt the device

to align its orientation with the silhouette of the


, which is displayed in



About the notes for the scan scene

Align and tilt the device to match the silhouette of the blue phone, which is displayed around the purple ring.


Scan Scene Notes 1

Orient the device towards the object


Scan scene note 2

Move farther away


Scanning scene caution statement 3

Move closer


Scan scene note 4


slowly around the object


Scan Scene Notes 5

After completing the scan,

select “Upload”


Upload scene

Once the

progress circle in the lower right is filled, the 3D model is complete


Upload Complete

Checking and Outputting the Completed Model

Once the 3D model is complete, you will see a thumbnail of the scene you shot


Selecting the thumbnail will bring up the following screen:


App version of the 3D scene


this scene you can

  • Playing the video of the 3D scene
  • Tap operation of the 3D scene
  • Tap operations on the 3D model
  • Outputting videos and 3D models


the video of the 3D

scene Select the leftmost of


three icons in the lower left corner of the screen


play the video of


3D scene


Scene icons

Tapping on

the 3D scene

Select the middle icon of the three icons in the lower left corner of the screen to tap around


3D scene.


Scene icons


on the 3D model

Select the rightmost of


three icons at the bottom left of the screen


tap on


3D model.
In this mode you can see the top of the 3D model and zoom in and out






動画や3Dモデルの出力 下にタブが


「Cinematic Render」を選


Web版のLumaAIのサインイン Sign in to the web version of Luma AI[/caption]

After successfully logging in, you will see thumbnails of the models you have taken so far


Thumbnail of the Web version

Checking and outputting the completed model

Select the model you want to check and

a 3D scene will appear


3D scene in the web version


basic operation is the same as on a smartphone, so I will skip the explanation.

Output video

If you want to save the video of the 3D scene, select the icon in the upper right corner.


Downloading the Web version of the video

Outputting a

3D model If you want to

output a 3D model, with the thumbnail of the model displayed, hover over the model you want to output and select the download icon:


Downloading a 3D model from the web version

Creating a 3D model from a video

With the

model thumbnail displayed, select “create new” in the upper right corner of the page to create a 3D model from your video


Creating a 3D model for the web version

Editing the 3D model

Edit the 3D model output by Luma AI and delete the parts you don’t need.
This time we will edit the OBJ file in Blender.

First, start Blender and import the OBJ file.


OBJ Import

After opening the OBJ file, change the mode from “Object Mode” to “Sculpt Mode


Select Sculpt Mode

Select “Box Trim” from the icons displayed on the left side


Select BoxTrim

Drag to select and delete unwanted portions of the imported 3D object


BoxTrim sample

Please note that in the video above, the deletion is done quickly, but in reality it takes much longer.

Once editing is complete, export the file in the file format of your choice.
This time we will export as an OBJ file


OBJ Export

When exporting as an OBJ file, set “Path Mode” to “Copy” as we also want to export the textures.


Exporting Textures


Sometimes when you do a Box Trim, strange colored objects appear.
If this happens, switch to “Edit Mode” so you can select faces, select all the parts you don’t want, and finally delete them with the “X” key.

Uploading to ST

YLY Once you

have finished Blender’s output, upload the 3D model to STYLY.

Create a STYLY account


How to upload to STYLY

How to upload to STYLY from Unity


For questions about STYLY, bug reports, and requests for improvements, please contact the STYLY FORUM

Certified (QA) by Shota Shawn Yoshizawa
Edited by SASAnishiki