Process your first 3DGS model

Blurry takes you from source images or a video to a processed, editable, and shareable 3D Gaussian Splatting (3DGS) model in one workflow. In this guide, you’ll upload a Capture, select the training parameters, start a processing run, and open the finished Scene.

Before you begin

Prepare your dataset of multiple images or a single video file. Don’t mix images and video in the same Capture. For the complete input requirements and supported file formats, read our Uploading a Capture documentation.

1. Upload a Capture

Create a new Capture from the Blurry dashboard and give it a descriptive name. Drag your images or video into the upload area, then wait for every source file to finish uploading. Review the files before continuing because Blurry locks them when the first processing run starts.

2. Process the Capture

Open the processing settings, keep the recommended defaults or customise the splat count, resolution, and spherical harmonics. Select Start processing and complete Stripe Checkout to authorise the payment. Blurry only charges you after the processing run finishes successfully.

After checkout, Blurry returns you to the Capture. Follow the run’s status under Existing runs. When the Scene is ready, select the completed run and use Open scene in the details sidebar.

Open the Scene in Blurry

Your processed model is now a Blurry Scene. Open it to review the result, adjust its presentation, add annotations or waypoints, and share or embed it wherever your audience needs to see it.

Continue with the editor documentation to learn how to navigate Scene contents, align the model, and refine it with masks. When the Scene is ready, you can also share it or embed it on a website.