Lister (Webcam & Recognition)

In this tab, you configure the camera for AI-supported part recognition. Since the Lister is the heart of your workflow, you should make sure that the optimal camera for your scans is selected here.

1. Webcam Settings

If you have multiple cameras connected to your workstation (e.g., a built-in laptop camera and an external USB camera above your sorting table), you can select the active camera here.

• Scan for cameras: Click this button to update the list of available cameras. Upon the very first click, your operating system (Windows/macOS) may ask for permission to access the camera.
• Active camera: Select the desired camera from the dropdown menu. The selection is saved immediately.

Camera not found or image remains black?

By default, webcams on Windows and macOS can usually only be used by one application at a time.

If your camera image in the Lister remains black or the camera doesn't even appear in the list, please check if another software (e.g., Zoom, Teams, OBS Studio, or an open browser tab) is currently blocking the camera. Close the other application and click "Scan for cameras" again.

---

2. Notes on Data Transmission & AI

Transparency is important to us. Since BLAPI-Core is a local application, we would like to briefly explain what data is transmitted to external services during part recognition:

Image Processing via Brickognize

The actual "intelligence" of the part recognition does not run on your computer but is handled via the excellent Brickognize API.

• As soon as you press the spacebar in the Lister, a photo is taken locally.
• To ensure rapid recognition in fractions of a second, BLAPI-Core compresses the image locally (to max. 800 pixels wide) and converts it into a highly compressed JPEG.
• Only this compressed image is sent to the Brickognize servers purely for analysis.

Color Feedback & Machine Learning

The AI is constantly learning. BLAPI-Core actively supports this learning process:

• If the AI recognizes a part correctly but gets the color wrong, you can manually correct the color in the Lister.
• As soon as you do this, BLAPI-Core sends anonymous feedback to Brickognize in the background ("The prediction was wrong, it was actually color X").
• Even if you simply confirm the suggested color (by adding the part), positive feedback is sent.

By using the software, you are actively helping to make the recognition models a little bit better and more precise for the entire LEGO community every single day!