20 layers between your
interview and cheating.
Every SViam interview is monitored, analyzed, and cryptographically sealed.
20
Integrity layers
500ms
Monitoring interval
0–100
Proctoring score
SHA-256
Tamper-evident hash
Before the interview
Identity and environment verified first
PIN authentication, a 15-second room scan, ambient audio baseline, and IP geolocation check — all verified before the candidate sees the first question. VPNs, cloud providers, and known proxy infrastructure are flagged automatically.
During the interview
9 layers running in parallel
Tab-switch detection, fullscreen enforcement, clipboard blocking, eye tracking, screen capture every 30 seconds, code snapshot timeline, typing cadence recording, real-time AST analysis, and dynamic integrity interventions.
When a signal fires, the AI interviewer asks a natural follow-up that verifies understanding — without ever accusing the candidate.
After the interview
Deep analysis catches what live monitoring cannot
Typing cadence classification (organic vs. transcription vs. paste-disguised), solution arrival curve analysis, AI code fingerprinting against known patterns, full audio forensics for secondary voices, and behavioral profiling that identifies inconsistencies between verbal explanation and code.
One number. Every signal weighted.
Every interview produces a proctoring score from 0 to 100.
80–100 Low|60–79 Medium|40–59 High|0–39 Critical
Audit Trail
Cryptographically sealed. Publicly verifiable.
Every completed interview generates a SHA-256 hash of the transcript, scorecard, session ID, and end timestamp. Anyone can verify a session hasn't been tampered with — no login required.
GET /interviews/sessions/{id}/verify-integrity
{
"tamper_detected": false,
"stored_hash": "a3f8c1...",
"computed_hash": "a3f8c1...",
"verified_at": "2026-05-11T..."
}
See it in action
Run a test interview and review the proctoring report yourself.