Is motion/subjects/objects detection performed locally in the camera or it's analized remotely in Wyze servers (via internet)?

I wonder if I depend on remote servers -and internet services- or just the camera internal software to get my notifications and run automation (rules).
Thank you!

If you have the setting toggled on for it to record events when there is motion then motion is determined locally on the camera. It compares the pixels from one frame against the pixels from 1 or more other recent frames and if there are enough differences (determined based on your sensitivity level settings), then it will decide there was motion.

Once motion is determined, then it will decide what to do from there:

  • If you have Cam Plus Lite, it will upload a short 12 second recording to the cloud and analyze whether or not there was a person in it. This is not done locally…at least not at the time of this writing. The AI Team recently told us in an AMA that they are looking to add more Edge/Local detection stuff in the future, but we don’t know much more than that they are working on some potential Edge (local) solutions.
  • If you don’t have Cam Plus Lite (nor Cam Plus), then it won’t upload a video recording to the internet, just a thumbnail image of what was onscreen when it detected motion (pixel changes)
  • If you have Cam Plus then it will upload a recording as long as it continues to detect sufficient motion and all of it will be analyzed by their servers while being uploaded to the cloud. Again, this is not done locally.
  • Regardless of the above options, if you have an SD card in the camera and have the option for it to record events only, then the camera will record the event onto the SD card as soon as it detects motion. It will record each full minute wherein any part of that minute had motion. That means, if motion started at 59 seconds into the minute, the entire previous 59 seconds that happened before this will be recorded along with the next second and if motion still continues into the next minute, then the entire next minute will be recorded, even if motion ends 2 seconds after that minute starts. So you will always have a recording of what happened just before motion was detected and what happened just after motion ended (again dependent on the sensitivity level settings to determine how much motion counts as sufficient motion).
    • If you have the SD card set to record on “Continuous Recording” then it doesn’t matter whether the camera detects “Motion” as it will record 24/7 to the SD card anyway, and when the SD card is full, it will just rewrite new recordings over the top of the oldest ones.

MOST automation rules are processed through the internet right now.

There are a few things that are done locally.

For example, some schedules and timers and such are local and will continue to execute even if there is no internet.

But currently no devices can really talk to other devices locally without internet (excluding Bluetooth devices talking directly to the phone or their local hub/base/gateway that connects them to the router…that communication is all local).

It is possible that some of this may change in the future as the VP of product and AI team said they are working on pushing more to be done locally instead of through the cloud. :+1:


Thank you very much, carverofchoice.
Awesome answer!
I have a clear vision now, and know what to expect if internet fails. (In addition to many other details on how this works)
Thank you,