Real-time AI patient
A 5-minute role-play with a live-audio AI patient. They listen for pauses, ask clarifying questions, and stay in character. No scripted branches, no canned replies — every conversation is shaped by what you say.
Features
Everything in one role-play —nothing wasted.
OET Live is the only AI Speaking coach that scores every OET criterion, surfaces the exact words you fluffed, and tells you which cues you missed — automatically, in under a minute.
Coming soon to iOSFree during early access
Core capabilities
Five things that make OET Live different.
9-criteria OET scoring
Every role-play is scored on the four linguistic criteria (intelligibility, fluency, appropriateness, resources of grammar and expression) and the five clinical communication criteria (relationship building, understanding the patient perspective, providing structure, information gathering, information giving). Plus a 0–500 overall and band letter.
Pronunciation diagnostics
Industry-grade pronunciation analysis scores every syllable for accuracy, completeness, and fluency. We surface the specific words you mispronounced, ranked by how much they would hurt an examiner — so you drill what matters.
Cue detection per task
Every OET role-card has 4–5 tasks. We verify each one against your transcript — hit, partial, miss — with the exact quote you used as evidence. No more guessing whether you covered all the cues.
Personalised recommender
A weighted recommender targets your two weakest criteria, suggests topics you haven't tried, and adjusts difficulty as your scores climb. Hit Recommended and you get a case engineered to move your weakest band.
The rubric, demystified
All 9 criteria scored on every role-play.
OET Speaking examiners grade against four linguistic criteria and five clinical-communication criteria. We mirror the public rubric and give you both the band and the why.
01 · Criterion
Intelligibility
How easy you are to understand at the sound and stress level.
02 · Criterion
Fluency
Speed, rhythm, hesitations, fillers.
03 · Criterion
Appropriateness of language
Vocabulary, register, hedging — does it fit a real consultation?
04 · Criterion
Resources of grammar and expression
Range and accuracy across complex structures.
05 · Criterion
Relationship building
Greeting, rapport, empathy, closing.
06 · Criterion
Understanding the patient perspective
Eliciting concerns, ideas, expectations.
07 · Criterion
Providing structure
Signposting, transitions, agenda-setting.
08 · Criterion
Information gathering
Open + closed questions, screening, summarising.
09 · Criterion
Information giving
Chunks, checks, jargon control.
Want this on your iPhone? TestFlight opens soon.
Join the waitlist and we'll email you the moment your build is ready. Free during early access.