The read-only XRInputSourceEvent property frame specifies an XRFrame object representing the event frame during which a WebXR user input occurred. This may thus be an event which occurred in the past rather than a current or impending event.

An XRFrame indicating the event frame at which the user input event described by the object took place.

The event frame does not correspond to a visual frame as is delivered to the frame rendering callback function (see Rendering and the WebXR frame rendering callback for details on the callback). Instead, the XRFrame specified by the frame property is a method to provide access to the getPose() method, which you can use to get the relative positions of the objects in the scene at the time the event occurred.

However, since the event frame isn't an animation frame, there is no viewer pose available to represent the viewer's current point of view; the results of calling getViewerPose() will be an XRViewerPose with an empty views list.

This code shows a handler for the selectstart event which gets the target ray's pose from the frame, mapping the pose representing the ray (event.inputSource.targetRaySpace) to the overall reference space myRefSpace.

Then, if the result isn't null, the target ray pose's transform is passed into a function called myCheckAndHandleHit() to see if the ray was pointing at anything when the select was triggered.

xrSession.onselectstart = (event) => {
  let targetRayPose = event.frame.getPose(
    event.inputSource.targetRaySpace,
    myRefSpace,
  );
  if (targetRayPose) {
    checkAndHandleHit(targetRayPose.transform);
  }
};