Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/Saurabh-07586/examplatform/llms.txt

Use this file to discover all available pages before exploring further.

Live Monitoring gives administrators a real-time view of every student session currently in progress across all active exams. Rather than waiting for a proctoring report after the fact, you can observe integrity signals as they happen, intervene when necessary, and force-submit any session that warrants immediate action. The page updates automatically so you never need to manually refresh.

The Live Monitoring Page

Navigate to AI Monitoring in the admin sidebar to open the Live Monitoring view.
  1. Select an active exam from the list on the left. Only exams with at least one session currently in progress are shown. If no exams are live, the page displays “No live exams right now.”
  2. Once you select an exam, a session table appears — one row per student who is currently taking that exam. The table auto-refreshes every 10 seconds to keep all values current.

Session Table Columns

ColumnDescription
SessionUnique identifier for the student’s exam session
Current QuestionThe question number the student is currently viewing
Time RemainingCountdown of time left in the student’s session, as reported by the server-side timer
Tab SwitchesThe cumulative number of times the student has navigated away from the exam tab
AI RiskThe AI-computed integrity risk level for this session: Normal, Elevated, or Critical
Last HeartbeatTimestamp of the most recent keep-alive signal received from the student’s browser
Force SubmitAction button to immediately submit the student’s exam with their current answers
If no students are currently taking the selected exam, the table displays “No students currently in progress for this exam.”
The session table refreshes automatically every 10 seconds. You do not need to reload the page. The Last Heartbeat column will update as each student’s browser checks in with the server.

AI Risk Levels

The AI Risk column reflects the overall integrity signal computed from all proctoring events recorded so far in the session.
No significant anomalies detected. The AI has not flagged any suspicious patterns. The student’s webcam feed, gaze tracking, and tab activity are all within expected parameters.

Force Submit

The Force Submit button in each session row allows an administrator to immediately end a student’s exam and record their answers up to that point as their final submission. When to use Force Submit:
  • The AI Risk level has escalated to Critical and the integrity of the session is no longer acceptable.
  • You have confirmed a technical issue on the student’s device that prevents a clean submission, and support has instructed you to close the session server-side.
  • The student is unresponsive and the Last Heartbeat has gone stale, indicating a disconnection that is not self-resolving.
Force Submit is immediate and irreversible. As soon as you click Force Submit, the student’s exam is closed and their answers at that exact moment are recorded as their final submission. The student cannot re-enter the exam or change any answers afterward. Use this action only when you are certain it is necessary.
If the Last Heartbeat timestamp is significantly stale — several minutes behind the current time — the student may have lost their internet connection or their browser may have crashed. In this case, the session will eventually time out automatically when the server-side timer expires. Force Submit can be used to close it immediately if waiting is not acceptable.
Live risk scores give you a real-time signal, but they are based on the events captured so far in the session. For the full integrity picture — including a timeline of every flagged event, all gaze-away and face-detection incidents, and the overall AI risk score summary — open the AI Proctoring Reports section after the exam concludes. Cross-referencing live observations with the post-exam report is the recommended workflow for any integrity investigation.

Build docs developers (and LLMs) love