kaiwalyaraut27-amt/heart-disease-predictor-project — reverse-engineered prompt

Build me a clean Streamlit web app called CardioScan AI that predicts heart disease risk from patient clinical data.

The user should be able to enter the common heart dataset values like age, sex, chest pain type, resting blood pressure, cholesterol, fasting blood sugar, ECG result, max heart rate, exercise angina, oldpeak, slope, vessels, and thalassemia. Use the existing trained model file and feature JSON to make the prediction, including the engineered flags like high cholesterol, high blood pressure, heart rate ratio, oldpeak risk, and age groups.

Show the predicted probability clearly with a friendly risk label, low, mild, moderate, high, or critical, plus a short suggested next step. Make the app look polished but simple, with helpful input labels, sensible ranges, and a strong disclaimer that this is educational only and not medical advice.

Please make sure it runs locally with streamlit run app.py and uses the files already in the project.