The GET People/Vehicle Counts endpoint allows users to programmatically retrieve detection counts for both people and vehicles across their camera fleet. The returned counts are pulled from detections made on thumbnails stored in our backend services.

Due to the architecture of this endpoint, it is important to understand how these objects counts are retrieved. Every 20 seconds, Verkada cameras send a thumbnail to our backend services with the bounding boxes and positions of each person and/or vehicle detected within the frame. If motion occurs within those 20 seconds, the camera will send an additional thumbnail including the metadata for each detected person and vehicle. This means that at most, Verkada cameras will send two thumbnails to our backend every 20 seconds.

For high traffic areas, the returned count may be an under representation of the actual count of people and vehicles present in the scene. The GET People/Vehicles Count should only be used to understand traffic patterns and trends as opposed to obtaining an exact count of people and/or vehicles.

To use this endpoint, users will need to input the camera_id from which they want to retrieve the object count from. Users also have the option to input a start_time and an end_time to query the count during a specific time range. If no time range is specified, the endpoint will retrieve the object counts for the past hour.

Note: Depending on the specified time range used to retrieve the people and vehicle counts, the returned JSON may be paginated. See the Pagination page for more information.

Below is an example of the returned JSON from this API endpoint. The associated Unix timestamps in seconds represents the time at which the thumbnail was generated.

200OK