AI viva practice
A viva is passed out loud, and practising silently does not prepare you for it. VivaVoice runs spoken exam stations with an AI coach: it questions you the way an examiner does, holds you to real exam timings, and returns feedback mapped to the official marksheet. One station is free to try.
You choose a station and the coach sets the task — an anatomy image, a clinical scenario, a communication encounter. You answer out loud against the station clock. The coach probes incomplete answers the way an examiner would, then scores the station and shows you the breakdown: where marks were won, where they were lost, and what a stronger answer covers.
Feedback maps to the marked domains, not general impressions: the content points the marksheet rewards, the structure and timing of your answers, the safety-critical points, and the communication criteria. Across attempts you see which stations and domains move, so revision time goes where the marks are.
A question bank tests whether you recognise the right answer; Part B tests whether you can deliver it, out loud, inside nine minutes. A live course gives you structured teaching, but a fixed number of hours, mostly spent watching — and one date. Spoken practice is the part neither format covers: here, every minute is you performing a station, as many times as a weak topic needs, at whatever hour you're free. Most candidates pair a question bank for content with VivaVoice for performance — they are doing different jobs.
MRCS Part B is live — a full mock OSCE. MRCP PACES opens in June 2026, built to the PACES23 format. Further exams follow.
Sai Chauhan, Founder · VivaVoice