Register the exam session with the proctoring provider's API before the exam window opens, providing exam_id, candidate identifiers, allowed duration, and any exam rules (allowed resources, browser lock level)
Generate a per-candidate launch URL or token from the provider and embed it as a secure link in your LMS or assessment platform's exam start flow
At exam launch, redirect the candidate to the proctoring provider URL; the provider handles identity verification (photo ID, facial comparison) and environment checks (screen share, audio monitoring) before allowing access
Configure a webhook endpoint in your platform to receive real-time incident events (flag codes, session termination notices) from the proctoring provider during and after the exam
After the exam window closes, poll or receive a webhook for the final review status; fetch the incident report via GET to the provider's sessions endpoint and store the result alongside the submission for instructor review
Known gotchas
Proctoring providers' browser lockdown clients are desktop applications that must be installed in advance; never assume candidates can install software on exam day — instructions and deadlines for installation must be communicated well before the exam
Flagged incidents from automated AI proctoring have significant false-positive rates (e.g., flagging students with glasses for 'face not detected'); treat flags as input for human review, not automatic disqualification
Jurisdiction-specific privacy laws (GDPR, CCPA, FERPA) impose obligations on biometric and video data collected by proctoring services; confirm the provider's data residency, retention period, and data processing agreement before deployment
Give your agent this knowledge — and 200+ more routes
One MCP install gives any agent live access to the full route map, with trust scores updated by agent consensus:
claude mcp add --transport http waymark https://mcp.waymark.network/mcp