NeckMoves guide

What can a camera-controlled movement game see? Four privacy checks

Before granting camera access, ask where frames are processed, what is uploaded or saved, when capture stops, and whether a usable alternative exists.

“Uses a camera” does not automatically mean “uploads video,” but people should not have to rely on a vague privacy promise. A trustworthy camera-controlled tool explains where frames are processed, what is retained, when capture ends, and what remains usable after permission is refused.

Short answer: check four things before granting access: whether processing happens on your device; whether video frames or face images leave it; whether the camera track stops when the activity ends; and whether you can refuse permission or use an alternative. A browser camera indicator shows device use, but it cannot by itself prove that a site sends nothing away.

What browser permission protects

The MDN documentation for getUserMedia explains that camera access is limited to secure contexts and requires user permission. Browsers must also provide indicators for active use and granted permission.

Those mechanisms stop a page from simply opening a camera without consent. They do not automatically decide what a site does with the stream after access is granted. The product remains responsible for its processing and network behaviour.

“On-device” should describe a data flow

NeckMoves uses MediaPipe Face Detector in the browser to obtain only the small amount of live direction information needed to control the game. The model and runtime are hosted by the site, and detection runs on the device. Google's MediaPipe Tasks privacy notice also states that input data is processed on device and is not sent to Google servers by MediaPipe.

NeckMoves does not upload video frames or save face images. Face landmarks and continuous angles do not enter account records or analytics events. Network policy permits only same-origin requests needed by the product.

Four questions before you allow access

  1. Where is processing performed? In the page, on the device, or on a remote server?
  2. What is retained? Video, screenshots, landmarks, direction values, or only a completed-round event?
  3. When does it stop? Are camera tracks closed after a round, on leaving the game, or when the tab closes?
  4. Can you say no? Is there a no-camera preview, keyboard path, or clear exit?

“We value your privacy” is not enough if a product cannot answer these operational questions.

What you can observe yourself

Confirm the browser or system camera indicator when granting access and check that it disappears after leaving. You can revoke the permission in browser site settings. A technical user can also inspect the Network panel for outgoing requests, but ordinary users should not need packet inspection to obtain basic trust.

The alternative in NeckMoves

If you prefer not to enable a camera, begin with the preview and a no-camera desk break. Camera access is an intentional user choice, not the price of having any movement option.

The strongest privacy result is not asking people to trust more. It is collecting less and making exit unmistakable.