Angelcam Car Detector
Pricing
Pay per usage
Go to Apify Store
Angelcam Car Detector
Checks Angelcam cameras for a parked car and reports presence, position and colour. Runs YOLOv8 locally - no external AI API keys.
Angelcam Car Detector
Pricing
Pay per usage
Checks Angelcam cameras for a parked car and reports presence, position and colour. Runs YOLOv8 locally - no external AI API keys.
Angelcam camera IDs to check. Leave empty and use 'Camera query' to resolve by name, or leave both empty to check every online camera.
Free-text camera reference (name, tag, location), resolved via the MCP find_camera tool. Ignored when Camera IDs are given.
'live' grabs a fresh frame from the MJPEG live stream (recommended - periodic snapshots can be 30+ minutes old). 'snapshot' uses the latest Angelcam periodic snapshot.
Optional parking-spot region as fractions of the frame: {"x":0.2,"y":0.4,"w":0.3,"h":0.4} (or pixels). inRegion is true when a detected vehicle box overlaps the spot by at least "Min region overlap".
Percent of the vehicle box that must lie inside the region to count as parked there.
YOLO confidence (percent) below which detections are ignored. Lower = more sensitive, more false positives.
Tile the region of interest and run the detector on each tile. Needed for cars far from the camera - a full 2560x1440 frame is squeezed into 640x640, shrinking distant cars below the detector limit. "auto" tiles whenever a region is set.
Tile edge in source pixels. Smaller = more zoom on distant cars, more tiles, slower.
Store the analysed JPEG in the run's key-value store and link it from the dataset row.
Base MCP endpoint of the Angelcam MCP server.
From npm run login. Falls back to ANGELCAM_CLIENT_ID env var.
