Veref Interview
The interview room built for integrity.
Browser-native video interviews with identity verification, real-time deepfake and voice-clone detection, AI co-pilot signals, and a live integrity score the recruiter can act on. No downloads. No plugins.
Step 1: Verify identity
Know who is about to join before they join.
Every candidate completes a 45-second identity check before the interview link unlocks. A government ID scan, a selfie liveness challenge, and a device fingerprint. If any of it fails, the interview never starts.
- Supports passports, driver licenses, and national IDs from 200+ countries
- Active liveness prevents photo, video, and mask replay attacks
- Biometric template stored, not raw images
- Device and browser fingerprint anchored to the session
Identity verification
Pre-interview · takes 45 seconds
- 01Government ID scan (passport, license, national ID)
- 02Selfie liveness challenge (active, not a photo)
- 03Biometric template stored as reference
- 04Device and browser fingerprint captured
Step 2: Lock down the room
A browser-native room purpose-built for trust.
Veref Interview runs in Chrome and Edge on desktop. No downloads, no plugins. Known deepfake drivers and virtual camera software are blocked at the system level before the candidate can even join.
- OBS Virtual Camera blocked
- ManyCam, XSplit blocked
- Deepfake driver signatures blocked
- Browser environment locked
Step 3: Detect in real time
Signals weighted into a score the recruiter trusts.
Veref runs deepfake detection on video frames, voice authenticity on the audio stream, and behavioral signals on gaze and latency. Findings are weighted into a live integrity score so recruiters always know where they stand.
- Real-time video deepfake scoring
- Voice anti-spoof and clone detection
- Gaze tracking for off-screen reading
- LLM-perplexity flags on the live transcript
- Deepfake video detectionClean
- Voice authenticityReal
- Virtual camera driverNone
- Continuous face match98.4%
- Gaze off-screen2.1s
- Response latency0.6s avg
Everything in Veref Interview
Six capabilities. One room.
Pre-interview identity
Government ID and selfie liveness before anyone joins the room.
Virtual camera block
Known deepfake drivers are detected at the system level and blocked on join.
Continuous face match
Every 30 seconds, the face on screen is matched back to the verified ID.
AI co-pilot signals
Response latency, gaze tracking, and LLM-perplexity flags on the live transcript.
Active challenges
One-click challenges the recruiter can trigger: head turn, hand pass, room pan.
Evidence report
Every session exports a PDF with flags, clips, and transcript timestamps.
See a live deepfake detection in a real interview.
We will run it on a real candidate flow, with a real test actor using real-world tools.