How it works

Four steps from setup to report

No terminal commands — the proctor only touches buttons in the website.

  1. 1

    Create the proctor account

    Signing up asks for a full name, username, email, ID type and a strong password. The proctor's own evidence folder is created at that exact moment.

    Create account I already have one
  2. 2

    Enroll the dashcams

    On the dashboard, press Scan for cameras. The system detects every camera physically connected to the PC — no URL or IP needed — then you give it a camera name and the room where it is installed.

    Go to the dashboard
  3. 3

    Start the livestream and auto-detection

    Press Start Livestream, then Start auto-detection. Every face in view becomes its own student frame with a live status — NORMAL, WARNING or FLAGGED — and suspicious frames are captured automatically with the reason behind the flag. Leaving the room? Press Lock monitoring.

    Start monitoring Read the features
  4. 4

    Review, search and export

    After the exam, open Logs & records. Every entry carries the date, time, status, student frame and reason. Search it, enlarge the screenshot, download the proof, or export the whole filtered list as CSV or PDF for the panel.

    Open logs & records

What the PC needs

Python 3, a webcam or USB dashcam, and the packages in requirements.txt. Start Smart Proctor.bat installs whatever is missing on the first run.

Camera permission

The browser asks for camera access the first time you scan for cameras or start a livestream — click Allow.

The AI models

OpenCV works out of the box. Installing ultralytics turns on YOLOv8 phone detection and mediapipe turns on head-turn detection — the dashboard shows which model is active.

Create a proctor account See who can open what