1) Ingest
Normalize readings, align timestamps, and validate ranges.
Data Flow
How NeuraGlu ingests your data, estimates insulin sensitivity, predicts near-term movement, and returns a precise carb suggestion in real time.
Inputs we use
You control what you share. These inputs improve accuracy and personalization.
- Blood Glucose — CGM or manual entries
- Insulin — dose & timing
- Activity — intensity & duration
- Weight & Age — baseline personalization
- Carb Response History — how your body reacts
Processing pipeline
2) Features
Trend, rate of change, insulin-on-board, recent activity markers.
3) Personalize
Estimate ISF per user from history and context.
4) Predict
Short-horizon glucose movement + hypo risk scoring.
Model & ISF
The model refines your insulin sensitivity factor (ISF) using history, context, and recent responses to carbs. This drives a precise carb dose suggestion, not a generic range.
- Adaptive ISF that updates with new data
- Confidence bounds to avoid over- or under-treating
- Real-time guardrails for safety
Output: precise carb suggestion
When hypo risk is detected, NeuraGlu returns a personalized carb amount with context (why, when, and how it was derived).
Suggested carbs
Personalized grams with an explanation of factors considered.
Timing
Immediate vs staged intake depending on predicted movement.
Follow-up
Optional recheck window and guardrails for retesting.
Adaptation
The system learns from your response to improve next time.
Privacy & security
Encryption in transit and at rest. Minimal data retention, user-controlled deletion, and transparent settings.
Get Early Access