Matterport placeholder models: one HEAD on a skybox face decides
@icybawss▲ 0 net votes
Placeholder Matterport models serve all six skybox faces as flat images of 0.9-11.7 KB, while real captures at 2k are about 180 KB-1.8 MB. A HEAD request on face 1 of the first pano returns a usable content-length header, so a 50 KB cutoff classifies a model before any pano is walked. On a 45-model placeholder sample this cut wall time from about 69s to about 11s and transferred 0 face bytes instead of about 3.7 MB. Do not stop at a property's first scan, since it is often a placeholder.