CrowdFlux

Count people with on-device AI. Track returning people, review session trends, and save annotated videos and reports. No calibration needed.

CrowdFlux turns your iPhone into a portable people-counting tool. Use the live camera or analyse an existing video, with detection and tracking processed on your device.

COUNT WITHOUT CALIBRATION Frame your scene and tap START. There is no counting line to position, no entrance or exit to define, and no distance calibration required.

FOLLOW PEOPLE IN MOTION Bounding boxes and temporary tracking numbers follow detected people as they move. Motion tracking and on-device appearance matching help keep the same session identity when a person returns. Ambiguous sightings remain Checking instead of increasing the identified estimate.

UNDERSTAND YOUR COUNTS Visible now shows confirmed people detected in the current view. Identified estimates distinct identities across the session, while Peak visible shows the highest simultaneous count. Unresolved sightings are reported separately and excluded from the identified estimate. A configurable visible-count threshold highlights when your selected level is reached.

SAVE VIDEO WITH TRACKING Optionally save silent video clips with bounding boxes, labels and statistics. Recording is off by default. Review or share clips from the saved session.

ORGANISE SESSIONS BY PLACE Add a venue name, city, country, time zone and optional filming location. Revisit session history, explore count charts and export CSV reports.

PROCESS ON YOUR IPHONE People detection and tracking run locally. Reports and optional recordings stay on your phone unless you choose to share them. Numerical appearance samples stay in device memory for the session to help match returning people. CrowdFlux does not recognise faces or infer gender. Map search and address lookup use Apple services and may require internet access.

IMPORTANT LIMITATIONS Counts describe the camera's view, not whole-venue occupancy. Identified is an estimate, not a verified number of unique visitors. Unresolved sightings can make the estimate low. People outside the frame or hidden behind others cannot be accounted for.

Low light, similar clothing, crowding, camera movement and longer absences can cause missed counts or duplicate tracks. Counts and threshold alerts must not be relied on for admission, legal capacity or safety decisions. Use independent checks and appropriate safety procedures.

Follow applicable privacy and recording requirements when capturing or sharing footage.