Voice Viva is in preparation. The examiner behaviour and voice quality checks are still being completed, so this mode is not yet promoted on the home page. Everything below describes how it will work. The text-based Part B OSCE mode is available now.

🎤 Voice Viva

Don’t just read orthopaedics. Say it out loud.

Reading an answer and saying one are different skills, and only one of them is examined. Press Voice in ChatGPT and talk through a station with an AI examiner.

Start practising Other practice modes

A station, start to finish

voice viva

YouStart a 9-minute Part B viva. Act as the examiner. Do not coach me until the station is complete.

ExaminerA 26-year-old presents four hours after a tibial shaft fracture treated in a backslab. He reports pain out of proportion to his injury. What is your immediate concern, and what will you do first?

YouCompartment syndrome. I would split the cast down to skin and reassess.

ExaminerYou have split the cast. The pain persists. What findings would you specifically elicit, and what would you not rely on?

Try this prompt

Copy one and paste it in.

Copy & start copies the prompt and opens OrthoSG in a new tab. Paste it in and you are examined.

Press Voice, then say: Start a 9-minute Part B viva. Act as the examiner. Do not coach me until the station is complete.

Viva me on the acutely swollen knee. One question at a time, examiner tone, no hints.

Run three back-to-back 5-minute vivas on trauma emergencies and feed back only at the end.

Open OrthoSG

What the examiner will and will not do

A viva is only useful if the examiner behaves like one.

It will

Present a concise clinical scenario
Ask one question at a time
Let you answer fully before responding
Probe incomplete answers
Stop cleanly at the end of the station
Give structured feedback afterwards
Name safety-critical omissions
Ask one near-transfer question

It will not

Interrupt you mid-answer
Coach you during the station
Hint before you commit
Accept a vague answer as complete
Reassure you instead of examining you
Certify your competence

If the examiner does any of the six on the right, that is a defect worth reporting. Use the feedback form and choose “Question or station quality”.

Before you start

Voice mode is a feature of ChatGPT, not of OrthoSG. Press the voice control, say your instruction, and the station begins. You can switch back to typing at any point.

Speech recognition is imperfect, particularly with drug names and eponyms. If a word is misheard, correct it and carry on — the examiner will not penalise a transcription error, but it is worth reporting if it happens repeatedly.

Voice-derived scores are formative only. They are not evidence of clinical competence.

Formative education — not certification of clinical competence. Do not enter identifiable patient information.

Start practising Give feedback