Clinical Simulator is an interactive training platform for hands-on medical practice. It combines guided simulation, real-time AI coaching, and structured post-session review to help learners build clinical confidence through active repetition rather than passive study.
Practice patient encounters from first question to final diagnosis.
- 3 patient cases across easy, medium, and hard difficulty
- Conduct history-taking via text or voice input
- AI patient responds naturally based on hidden clinical details, personality traits, and speech patterns
- Real-time coverage tracking shows how much of the clinical picture has been explored
- Clinical support tools available during the session: interviewing tips, critical gap alerts, and optional diagnosis reveal
- Submit a final diagnosis with clinical reasoning
- Receive a structured evaluation across 6 dimensions:
- Information Gathering (25%) — completeness of history, PMH, drug history, social history, review of systems
- Clinical Reasoning (25%) — hypothesis-driven questioning and differential diagnosis
- Diagnostic Accuracy (20%) — correct diagnosis and clinical understanding
- Communication (15%) — empathy, rapport, and patient-centered approach
- Efficiency (10%) — information yield per turn, no redundant questions
- Safety (5%) — red flag recognition and appropriate urgency
- Competency levels from Novice to Expert based on weighted score
- Beginner mode blocks early submission until 40% coverage is reached
Practice emergency resuscitation with camera-based motion tracking and real-time feedback.
- 3 scenarios with progressive difficulty:
- Hands-Only CPR (Beginner) — continuous compressions, no protocol steps
- Adult BLS Sequence (Intermediate) — full 30:2 protocol with scene safety, responsiveness check, 911 call, breathing check, compressions, and rescue breaths
- Advanced Emergency Response (Advanced) — 30:2 under pressure with multi-cycle endurance
- Camera-centric layout: webcam feed occupies the primary view with pose skeleton overlay
- MediaPipe pose detection tracks 33 body landmarks at 30 fps
- Compression detection via wrist motion peak analysis with dynamic threshold and debounce
- BLS phase guidance: step-by-step overlays walk through Scene Safety, Check Response, Call 911, Check Breathing before compressions begin (Intermediate/Advanced only)
- Ventilation phase: confirm 2 rescue breaths between compression sets in 30:2 mode
- 2-minute cycle break with rescuer switch prompt and cycle statistics
- Real-time feedback on the camera overlay: timer, compression count, and coaching instruction bar
- ActionStatusCard shows live form assessment: hands visible, arms straight, hands centered, full recoil
- Metronome at 110 BPM and voice coach available during training
- Post-session evaluation across 7 dimensions:
- Rhythm (35%) — compression rate adherence to 100-120 CPM target
- Form (25%) — hand visibility, arm straightness, and centering
- Depth Proxy (10%) — estimated compression depth from motion amplitude
- Recoil (10%) — complete chest return between compressions
- Compression Fraction (10%) — active compression time vs. total elapsed
- Rate Consistency (5%) — stability of inter-compression timing
- Readiness (5%) — BLS checklist completion
- Select a training module from the home screen
- Brief — review the scenario, scoring criteria, and setup checklist
- Train — live simulation with real-time cues and support
- Review — structured evaluation with score breakdown, strengths, gaps, and next-step recommendations
- Repeat — retry the same scenario or try a harder one
Session history is saved locally and displayed on the home screen for progress tracking.
- Camera and conversation take center stage — controls stay minimal and peripheral
- Phase-based progression with clear visual indicators
- Strong guidance for first-time users, efficient for repeat practice
- All warnings and confirmations are inline UI elements, not browser dialogs
- Feedback is specific enough to improve the next attempt
- Medical students practicing clinical reasoning and patient interaction
- Healthcare trainees learning BLS/CPR procedural skills
- Simulation-based training programs looking for scalable practice tools
- Teams exploring AI-driven, multimodal clinical education
npm install
npm run devRequires a modern browser with webcam access for CPR training. AI features require API keys configured in the settings panel (Gemini or OpenAI).
##Links 飞书：https://my.feishu.cn/docx/R0DrdkDcqoJnxgx6RlxcudOyn3c?from=from_copylink web link: https://clinical-simulator-zrwk.onrender.com/