A flag is not a verdict.
Tab switches, screen and camera recording, where someone is looking and whether they're speaking. It all lands on a live timeline your interviewer can watch. Nothing gets blocked and nothing gets scored automatically. These are flags, not verdicts.
- 60s
- away before the session ends
- Flagged
- for a person to read, never auto-scored
- Told first
- nothing watched without consent
What it actually watches
Everything it notices goes straight to your interviewer as it happens. None of it is ever judged automatically.
Tab switch — screenshot on the spot
Leave the tab and a screenshot is taken straight away. They get a nudge after three seconds, then another every twenty. At sixty seconds the session simply ends. That's a plain time limit, not the AI deciding anything about them.
Screen recording
The shared screen is recorded for the whole session, and if the share drops you'll know within five seconds. Copy-paste and the usual developer shortcuts are noted as they happen. Nothing is blocked, and nobody is forced into full-screen.
Camera recording
Video and audio are recorded throughout. If the camera or mic drops, that's flagged within five seconds, and an obviously faked or out-of-sync feed shows up too, because the mouth has to match the voice.
Where they are looking, and whether they are speaking
Head turns, gaze direction and lip movement are read off the camera. On the timeline they sit there as what they are, which is a picture of someone's attention. They never add up to a cheating score.
One live timeline
Every flag reaches your interviewer as it happens, then stays on a permanent timeline attached to the scorecard. A person reads it and decides what it meant.
Evidence, not accusations
A person decides, every time.
Nobody is ever failed, rejected or marked down by the proctoring itself. It tells you what happened and your interviewer reads it, because the same flag can mean someone cheating or a courier at the door, and only a person can tell those apart.
The candidate knows exactly what is being watched.
Five consent screens spell out every one of them before the interview opens: recording, AI analysis, monitoring of their surroundings and attention, session integrity, and what happens to the data. Nothing gets captured that they haven't already been shown in plain language.
Every flag stays on the record.
Each one carries how serious it was, when it happened and what was done about it, and it stays attached to the scorecard. If a hiring decision is ever questioned, that timeline is your answer, and it protects the candidate just as much as it protects you.
Before any of this starts
- Recording isn't optional once the interview starts. Five consent screens have to be accepted first, and nothing begins until they are.
- Camera and microphone get checked before the session opens, rather than discovered halfway through.
- Every kind of flag carries how serious it was and what was done about it, so the timeline reads as a record rather than a score.
See it run on one of your own roles.
Free to try. No credit card, and no credit spent until you ask for work.