Personalized glucose predictions powered by machine learning trained on real CGM data
Enter your data and click predict to see your personalized glucose response
Trained on 100+ meal responses from CGM data with varying macronutrients and personal health parameters
Uses a simplified gradient boosting approach to learn non-linear relationships between meals and glucose
Adapts predictions based on your BMI, HbA1c, and age to provide individualized glucose forecasts
Model achieves ~15 mg/dL average error on test data, continuously improving with more data
Here's a glimpse of the data used to train this model:
| Carbs | Protein | Fat | Fiber | BMI | HbA1c | Peak Glucose |
|---|