§ FAQ
Frequently
asked questions.
About the product
- What is GR:eat?
- GR:eat is a metabolic companion app. You log a meal — by photograph or by description — and it predicts how your body will respond, before the spike. Peak, rise, time to peak, a two-hour AUC, a 0–100 meal score, and up to three personalised recommendations.
- How accurate is the prediction?
- Predictions come from a two-tier model: a cohort regressor for the population and a per-user personalisation layer that engages from your fifth CGM-backed meal. A formal accuracy validation study is in scope for 2026; until that publishes, treat predictions as informative, not prescriptive.
- Do I need a continuous glucose monitor?
- No. The cohort baseline gives you predictions from day one. A CGM — Dexcom, FreeStyle Libre, Medtronic, others — closes the loop: every meal becomes a training example for your personal model, and the predictions improve at your body over time.
- Which platforms is the app available on?
- iOS and Android. The app integrates with HealthKit on iOS and Health Connect on Android.
About the prediction stack
- What food data is the app using?
- The McCance & Widdowson Composition of Foods Integrated Dataset 2021 (CoFID) — the UK national food composition reference — is wired in as the canonical macronutrient source. The AI vision pipeline returns a meal label, the app canonicalises that label to a CoFID entry, and the macros come from there.
- Where does the glycemic-index value come from?
- Every meal prediction now ships with a GI value, a band classification (low / medium / high) and a source tag. When a measured GI is available from a licensed dataset we use it; otherwise a proprietary heuristic estimator computes a GI from the meal's sugar fraction, fibre density, fat density and protein density relative to carbohydrate content.
- Does the app see my photo?
- Meal photographs are uploaded to render your history and to drive the AI vision pipeline that identifies food classes and estimates portion size. They are stored in Firebase Storage under your account. You can delete them at any time.
About privacy
- Where is my data stored?
- Firebase Authentication, Firestore (per-user meal records, predictions, settings) and Firebase Storage (meal images). HealthKit and Health Connect data we read stays on-device except for the specific transmissions documented in our data-flow specification — available on request.
- Do you sell my data?
- No. We do not sell user data and do not run third-party advertising on it. De-identified aggregates are used to improve the cohort model.
- Can I delete my account?
- Yes, in-app. Profile, meals, predictions and connection records are removed. CGM data forwarded to us via Terra is governed by Terra's retention terms in parallel.
About the company
- Is GR:eat a medical device?
- Not yet. GR:eat is provided today as a wellness companion. We are building the documentation, traceability and architecture appropriate to a regulated medical device. A submission is in scope for 2026.
- What is the company doing about regulation?
- Three frameworks are engaged: IEC 62304 (software lifecycle, Class B candidate), ISO 14971 (risk management — workstream open), and the FDA's Human Factors Engineering guidance (use-related risk analysis in scope). See the /evidence page for the current state.
- Where is the company based?
- Glucose Republic Ltd was founded in Cardiff, UK in 2018 and is a distributed company today — Cardiff, UK and Perth, Western Australia. There is no single headquarters.
For partners
- Do you offer an API?
- Yes — prediction-as-a-service is one of three partner offerings. A live demo of the prediction algorithm is available at grmo-mvp-staging.web.app. For commercial access, email partners@glucoserepublic.com.
- Can we white-label GR:eat?
- Yes — the existing app, in your colours, on your distribution. Two-tier prediction, CGM feedback loop, HealthKit and Health Connect already wired. Staged rollout is supported.
- What about the dataset?
- A growing free-living glycemic dataset paired with continuous-glucose telemetry and meal composition is available to academic and commercial partners under license. Consented, de-identified, partner access on request.
Question we didn't answer? Email hello@glucoserepublic.com and we'll either reply or add it here.