Here is exactly what the copilot does in each situation — and what you see on screen.
The copilot hears the meeting and your microphone separately, recognises each question once the interviewer finishes, and streams a one-minute answer built from your resume — while you keep eye contact.
Every interviewer line is translated as it is spoken (English ↔ Chinese), and hints come in both languages, so you understand precisely and answer fluently.
Press ⌥S — or let auto-capture do it — and the copilot reads the coding problem, diagram or error on your screen and explains a solution with its trade-offs.
Upload your preparation document and every prepared question is extracted. A spoken AI interviewer asks them in order, at random or improvises, follows up once, and never repeats itself.
Already talking? Expand answer continues from what you just said with the most valuable facts you have not mentioned yet, in under a minute.
Each session ends with a debrief: every question reviewed, what was missing, a stronger spoken version — and the recording to replay.