RealityScan command-line guide: From photos to a 3D model

What will you learn in this guide?

RealityScan’s command-line interface lets you sequence some application operations using commands and run recurring tasks with a batch file. In this guide, you’ll add photos of an architectural space to a project, align them, save the project, run processing steps configured in the application, then simplify and export a model. The commands use example folder names; adjust them to match your actual working folders.

This method does not guarantee that every operation will be automated. For example, the -start command runs the operations configured for the Start button in RealityScan. The application settings and command-line steps therefore need to be considered together.

Requirements

The official documentation demonstrates command-line examples using RealityScan.exe. The steps below are for Windows Command Prompt and .bat files. The documentation does not specify a required version, hardware tier, account type or pricing information, so check your installation and license terms separately before starting a project.

The example uses C:\RS\Images and C:\RS\Interior.rsproj for the photo folder and project file. Create these folders on your computer and place the images in the appropriate location. If your file paths contain spaces, verify how the parameters are parsed in your installation; the examples use paths without spaces.

Adding the RealityScan folder to the system PATH reduces the need to repeatedly enter the full application path in commands. The official documentation notes that the PATH line below applies only to the currently open command-line session.

Numbered application steps

1. Align photos in draft mode

First, add the images and run a quick draft alignment. This example adds the application folder to PATH, then imports photos from a folder and saves the project:

set PATH=%PATH%;C:\Program Files\Epic Games\RealityScan\
RealityScan.exe -addFolder C:\RS\Images -save C:\RS\Interior.rsproj -draft -quit

-draft runs alignment in draft mode. This initial check helps you see whether the photo set comes together. Rather than moving on to full alignment without reviewing the result, it’s safer to open the project in the application and check it first.

2. Include subfolders if needed

If the images are spread across subdirectories of the main folder, enable subdirectory scanning before adding them:

RealityScan.exe -set "appIncSubdirs=true" -addFolder C:\RS\Images -save C:\RS\Interior.rsproj -draft -quit

This setting makes the -addFolder command include subfolders. If the images are already in a single folder, you don’t need this additional setting. Rather than running both examples consecutively on the same project, choose one workflow to match your photo folder structure.

3. Run full alignment

After reviewing the draft result, reload the saved project and start full alignment:

RealityScan.exe -load C:\RS\Interior.rsproj -align -save C:\RS\Interior.rsproj -quit

The commands run sequentially: the project opens, alignment runs, and the result is saved to the same project file. Command-line parameters work in a way similar to operations performed in the application; the application can open normally and allow user interaction during processing. The -headless option is also documented for hiding the interface, but keeping the result visible may make it easier to check during initial tests.

4. Run the processing workflow configured in the application

After alignment, configure the operations required for model generation under RealityScan’s Start button. According to the official documentation, -start runs the processes configured for that button. Don’t assume that adding a command alone determines which modeling steps will be performed.

RealityScan.exe -load C:\RS\Interior.rsproj -start -save C:\RS\Interior.rsproj -quit

Check the options used during processing in the application, based on your project and goals. The command line makes it easier to rerun this predefined workflow; it does not guarantee that the settings are appropriate.

5. Simplify and export the model

If you know the model’s name in the project, select it with -selectModel, set the target triangle count with -simplify, and export it with -exportSelectedModel. The example below uses the name Model1 and a target of 1,000,000 triangles; this is an illustration of the command, not a quality recommendation.

RealityScan.exe -load C:\RS\Interior.rsproj -selectModel Model1 -simplify 1000000 -exportSelectedModel C:\RS\Interior.obj -save C:\RS\Interior.rsproj -quit

Replace the model name with its actual name in the project. If you need more detailed control over simplification and export settings, you can pass a params.xml file exported from the relevant tools in the application as an optional parameter to the commands. Which settings are used depends on how the file was prepared in the application.

Common mistakes

• Calling RealityScan.exe without running commands from the application’s installation folder can result in the executable not being found. Use the full application path or add the PATH line to the active command session.

• Don’t assume that images in subfolders will be added automatically. If you need this behavior, use the appIncSubdirs setting.

• Don’t treat draft alignment as a finished result. -draft and -align are different alignment commands; check the project after the draft stage.

• Don’t guess which operations the -start command will run. Configure the Start button’s settings in the application beforehand.

• Selecting a model by the wrong name or entering an incorrect project path can prevent the command workflow from using the expected file. Make sure the project and model names match those in the application.

Implications for architecture, interiors and archviz workflows

Teams working with models created from photos of spaces or buildings can use the command line for projects that repeat the same import and alignment steps. Simplification and export are also among the documented tools, making it possible to set up a workflow for preparing photo-derived models at different stages. This is not a replacement for architectural decision-making; it’s a tool for organizing photo-based data in a project and repeating specific operations.

For projects involving multiple people in an office, it’s important to keep folder structures, project names and processing settings in the application consistent. The RealityScan documentation does not provide information here about compatibility with specific archviz software, required hardware or license costs. Test how the exported file will be used in the target application, how hardware affects processing time and what the license terms are, using actual project data.

Next steps

Once the basic workflow is working, you can explore other project commands, such as adding images with an image list and loading and saving projects. Alignment options include exporting and reusing XMP camera metadata. The model tools documentation also covers cleaning, closing holes, calculating textures and other export operations. Testing each new command on a small copy of the project first can help reduce the risk of overwriting existing work.

Sources and licensing

This guide is based on the project and image commands, alignment commands and model tools sections of Epic Games’ RealityScan command-line documentation. The sources are marked as freely licensed; the content has been adapted into Turkish, and the examples have been edited for this guide.