AI virtual-patient simulation

Interview the patient.
Then face the examiner.

Work a clinical case at your own pace with an AI patient who answers only what you ask — or sit the timed oral exam on the same case. Both come back scored, with written feedback.

New accounts start inactive — an administrator or supervisor activates yours before your first sign-in.

History Taking Example
You
What brought you in today?
Patient
I've had a burning pain in my chest since last night.
You
Does anything make it worse?
Patient
The whole history is marked out of 100
against an editable rubric, with written feedback
One question at a time

The patient answers only what you ask, and never names their own diagnosis.

Up to seven stages

Your supervisor switches stages on per case, from history to case discussion.

Type or dictate

Answer with the keyboard, or use your browser's speech recognition.

Feedback, not just a number

Every marked answer comes back with a written explanation.

Two paths, one patient

Practise the case. Or be examined on it.

A case can carry both, and they are independent — you can sit the oral exam without having started the training walkthrough.

Path A · Training

Self-paced, untimed

Walk the whole case and find out where your reasoning breaks.

  • Interview the AI patient in free text — ask anything.
  • Choose examinations and investigations, then read the findings.
  • Commit to a diagnosis, a management plan and medications.
  • Progress is saved as you go — pick the case back up later.
Path B · Oral exam (SOE)

Timed, marked, one sitting

A structured oral examination led by an AI examiner.

  • Questions your supervisor wrote, served one at a time in stage order.
  • Each question carries its own countdown and its own marks.
  • Graded against a model answer you never see.
  • Retaking requires supervisor approval.
The training path

One case, stage by stage

A supervisor decides which stages a case uses, so not every case runs all seven. Whatever is switched on, each stage you finish is scored and explained.

Your overall score is the unweighted average of the stages you completed, graded A–E.

  1. 1
    History Taking
    Interview the AI patient; the conversation is marked out of 100.
  2. 2
    Clinical Exam
    Pick body part and tool; read the case-specific findings.
  3. 3
    Investigation
    Order the right tests and interpret the reports.
  4. 4
    Diagnosis
    Write your diagnosis; the AI marks accuracy and reasoning.
  5. 5
    Management Plan
    Set out your plan against the case's expected plan.
  6. 6
    Medications
    Prescribe from the catalogue — wrong picks carry penalties.
  7. 7
    Case Discussion
    Answer the supervisor's written questions and MCQs.
The other door

The Interactive Oral Exam

An AI examiner puts your supervisor's questions to you one at a time, in stage order. You answer by typing or by speaking; the clock runs on every question, and when it stops your answer is submitted as it stands.

Timed per question
Each question has its own limit and its own marks.
Read aloud
Each question can be spoken to you — mute it any time.
Model answer stays hidden
You get your marks and feedback — never the ideal answer.
Retakes are controlled
A supervisor must approve before you can sit it again.
Examiner · Diagnosis 0:41
Given the history and the chest film, what is your leading diagnosis — and why?
Community-acquired pneumonia: fever, productive cough and focal crackles, with lobar consolidation on the film.
Type or dictate marks per question

Illustration of the exam screen

Marking

You never just get a number

Some answers are judged by the AI; others are scored by exact comparison. Both come back explained.

Marked by the AI
  • The whole history-taking conversation, against a rubric.
  • Your written diagnosis — accuracy, completeness, reasoning, terminology.
  • Your management plan, against the case's expected plan.
  • Free-text answers, and every oral-exam answer.
  • Your differential list — returned as correct, missing or unlikely.
Scored automatically
  • Multiple-choice questions.
  • Which examinations and investigations you selected.
  • Which medications you prescribed — wrong picks carry fixed penalties.
  • Your overall training score: an unweighted average, graded A–E.
For supervisors & examiners

Built for the people writing the cases

Author a case — or let the AI draft it

Write the history, findings, investigations, medications, questions and MCQs yourself, or describe the case in one line and review an AI-generated draft before you save it.

Set the oral exam

For each examiner question set the stage, the model answer used only for grading, a time limit and the marks it carries. The AI can draft a full set for you to edit.

Follow every student

See training progress and oral-exam results per student, record a final evaluation, approve retake requests, and export to Excel.

Your institution chooses the AI
OpenAI Anthropic Claude Google Gemini OpenRouter OpenCode Zen Ollama, running locally

An administrator picks the provider and model, and every prompt the platform uses is an editable template — no redeploy needed.

Straight answers

What this is — and what it isn't

A written clinical simulator for teaching and assessment. No real patients are involved, and nothing here replaces supervised clinical practice.

No video or voice calls. Despite the name, consultations are written — there is no live call with anyone.
Dictation uses your browser. Voice input relies on built-in speech recognition, so it works best in Chrome or Edge.
The interface is in English. You can dictate in Arabic, but the screens are English and left-to-right.
No bundled case library. Your department writes the cases it wants to teach.
Accounts are approved by hand. Register as a student or a supervisor, then wait for activation — no email is sent.

Your patient is waiting.

Create an account, or sign in and open your first case.

Umm Al-Qura TeleHealth — virtual-patient simulation © 2026 · Privacy