Pick a scenario and an examiner. The AI examiner asks one question at a time, listens to your answer, and probes weak spots — same way a real CRNA program panel does. End the session and you'll get a scored rubric.
1. Pick a scenario
2. Pick your examiner
Voice loop status: the audio capture + speech-to-text + AI examiner reply + text-to-speech pipeline runs locally on the GPU. The first version uses text-only input/output; voice will be enabled once the /crna/voice/turn endpoint is wired (next session).