The AI that grades your real role-plays.
Record a full consultation with your study partner. Stop. Get an RCGP-style grade — three domains, every criterion, four skill scores — in 60–90 seconds.
You already role-play with friends. Now you get the feedback too.
Most candidates practise with a partner and leave guessing. AI Marking turns that same session into structured, RCGP-style feedback.
✕ Practice without marking
- "Yeah, that went okay I think?"
- No structured feedback on the three domains
- No record of what you actually said
- Nobody catches the cue you missed
✓ With AI Marking
- Graded pass / borderline / fail per domain
- Criteria marked achieved, partial, or missed
- Direct quotes showing what worked
- Every attempt saved to progress dashboard
Start from the case library
Head to the case library and choose any case. Cases are organised by clinical area and difficulty — pick something you're working on or just dive in at random.
- 12 clinical areas, each with 5–25 cases
- Progress indicators per area so you can see what's left
- Difficulty filters — tune cases to where you are right now
Hit Mark with AI on any case
Every case has a green Mark with AI button at the bottom of the case page. One tap takes you through to the marking screen with the right case already loaded — no setup, no fiddling.
- Case auto-loaded — no reconfiguring
- 12-minute timer starts only when you're ready
- 1 credit charged after you confirm grading
Record with a partner. Get graded.
Grab a study partner, hit record, and run the consultation for up to 12 minutes. Phone or laptop picks up both voices. Stop when you're done — your grade arrives in 60–90 seconds.
- Both voices picked up, auto-diarised
- Exam-pacing timer — amber at 8 min, red at 11
- Audio discarded after grading; transcript stays
What the marking actually tells you
Every piece of feedback is criterion-tagged and backed by something you actually said. Not platitudes.
Domains
Data Gathering · Clinical Management · Relating to Others — each banded RCGP-style.
Criteria tagged
Each marking criterion marked achieved, partial or missed, with a specific reason.
Skill scores (0–5)
Cue handling · Explanation · ICE · Empathy — the four that decide borderline passes.
Direct quotes
When we flag a missed cue, we quote what the patient said. No generic coaching.
A result page you'll want to read
Overall band up top, three-domain breakdown, four skill scores, then criterion-level feedback — all on one page, all saved to your dashboard for later review.
- Every attempt versioned against the case
- Compare skill scores across attempts over time
- Exportable as PDF for your educator
Everything you need to know
Short answers to the questions we get most. Anything missing? Drop us a note.
Do I need a study partner?
AI Marking needs two real voices. Practising alone? Use our AI Patient bot — it role-plays the patient and grades you.
How different is it from AI Patient?
AI Patient replaces your study partner. AI Marking grades a session with a real partner. Most candidates use both.
How long does grading take?
Transcription takes seconds. Grading takes 60–90 seconds. Close the tab — results save to your dashboard.
Is my recording stored?
Audio is discarded after transcription. Transcript and grade are kept. Don't use real patient details.
What does it cost?
1 credit per consultation. Pay-as-you-go, no subscription. Credit packs on the portal page.
Grab a partner. Record a case. Get graded.
Your first calibrated feedback on a real role-play is one tap away. You'll know within 90 seconds.