How video analysis works
Video coaching is optional. Here's exactly what happens when you turn it on.
Step 1
Your camera shows you on screen
The video stays inside your browser tab. It isn't recorded to a file, and it isn't streamed anywhere.
Step 2
Your device analyses it, frame by frame
A face-tracking model downloaded to your browser estimates where your face is and which way it's turned. This runs entirely on your own device.
Step 3
Only a few summary numbers are kept
When you stop recording, the camera switches off and we keep only numbers — like the share of time your face was in frame. No image or frame is ever sent.
Step 4
Your audio is transcribed, then discarded
Your voice is sent once to be transcribed into text. We keep the text and speaking metrics such as pace. The audio itself is not stored.
What we keep
- The written transcript of your answer
- Speaking pace (words per minute) and filler-word count
- Face-in-frame %, gaze-toward-camera %, and a head-movement score
- Your score and written feedback
What we never keep
- Any video, image or single frame
- Any audio recording
- Judgements about personality, honesty, emotion or confidence
The visual measurements are rough coaching signals, affected by lighting and camera angle — treat them as a nudge, not a precise score. You can delete everything at any time from Settings.