Skip to main content
Computer vision

Retail analytics that
identifies no one.

On-site computer vision detects and follows people as anonymous shapes — never faces — turning video into visits, journeys, how long customers stay, and queues.

  • No facial recognition
  • Anonymous by design
  • Processed on-site
  • Counts · journeys · queues
ARCHITECTURE

Video stays in your store. Analytics can travel.

All video is processed on-site, inside your store, and is never uploaded. Sending anonymous analytics to a secure cloud is a reporting choice, not a requirement, so a single store can operate on its own.

IN THE STORE IP · ONVIF/RTSP Smart camera NVR stream CountPort edge runtime counting · tracking · journeys · zones · queues Local dashboard Raw video stays on-site metrics only anonymous · aggregated REPORTING · ACROSS LOCATIONS Portfolio dashboard API / exports Secure cloud

Runs in your store by default

The software runs on a computer inside your store. A single store can keep operating on its own, without depending on an internet connection to the cloud.

Flexible hardware

The software can run on a back-office PC, a Mac Mini, a small dedicated computer, or supported smart cameras. You can choose the option that fits your store.

Privacy boundaries

CountPort is built around anonymous analytics. It does not use facial recognition and does not identify individuals. It does not estimate age, gender, or emotion.

WHY IT MATTERS

Understand the whole store, not just the door.

A counter at the door can tell you that someone entered. A system that sees the whole store can also show whether that visitor moved through an area, returned to a display, joined a line, or left quickly.

  • Exclude staff so your visitor numbers reflect customers, not your own team.
  • Measure how long customers stay and how busy each area gets through the day.
  • Use the same processing for color maps showing where customers spend the most time, store areas, customer routes, and lines.
Entrance Checkout 1anonymous shopper journey
PRIVACY

Built around summary retail figures.

CountPort is designed for everyday retail measures: how many people visit, the routes they take, how long they stay, lines, store areas, and the share of visitors who make a purchase. It does not use facial recognition and does not identify individuals to produce these figures.

What is processed

Video from your cameras is analyzed on-site to count people, follow movement, measure how long customers stay, and understand how the store is used.

What is stored

Only analytics records, such as visitor counts, store areas, timestamps, and summary figures. How long these records are kept depends on your plan and setup.

What is sent to the cloud

Only reporting data, so that you can view results across your locations. All video is processed on-site, inside your store, and is never uploaded.

Processing video on-site, without facial recognition or biometric identification, reduces privacy risk and simplifies data-protection review. For the full policy wording and privacy details, read Legal & Privacy.

See it in one of your stores.

Begin with the cameras you already have and confirm the setup in a single store before planning a wider rollout.

Request a demo