CarbCam: About

About CarbCam

CarbCam is a photo-based app for carb estimation for people with diabetes, focused on accuracy, privacy, and a clear, fast workflow for everyday use.

The app is available on iOS and Android and is functionally identical across both stores. The shared core is written as a Kotlin Multiplatform Module, while the UI is native (SwiftUI on iOS, Jetpack Compose on Android).

The Idea

CarbCam was born from the 10be ecosystem, originally to complement the hosting for Nightscout on ns.10be.de with a user-friendly photo tool.

Estimating carbohydrates at the dinner table is one of the most frequently repeated micro-tasks in daily diabetes management. One camera, one tap, one well-founded estimate, that is the idea.

CarbCam: Scan result (Android)
Back to top

Platform & Architecture

iOS

  • iOS 18+
  • SwiftUI, Swift 5.9+

Android

  • Android 8+
  • Jetpack Compose, Kotlin

Shared Module

Kotlin Multiplatform, nutrition logic, Nightscout wire, I18n, validation, formula calculation. Single source of truth.

Server

ns.10be.de: analysis proxy to several AI providers (Anthropic Claude, Google Gemini, OpenAI, xAI Grok, Mistral, OpenRouter) + Azure OpenAI and Zhipu GLM with BYOK, feedback system, update distribution, status page. The same server that also runs the Managed Nightscout Hosting.

Back to top

Open-Source Libraries Used

CarbCam uses these libraries, many thanks to the maintainers:

iOS-specific

  • TOCropViewController MIT: Image cropping
  • Apple Vision, Barcode detection & local image analysis

Android-specific

  • Jetpack Compose Apache 2.0: UI toolkit
  • Room Apache 2.0: Local DB
  • Google Maps Compose Apache 2.0: Maps & Location
  • Coil Apache 2.0: Image loading
  • Retrofit / OkHttp Apache 2.0: HTTP (Square)
  • ML Kit Barcode Apache 2.0: Barcode scanner (Google)

Shared (Kotlin Multiplatform)

  • Kotlin & Coroutines Apache 2.0: JetBrains
  • Ktor Apache 2.0: HTTP for KMP
  • kotlinx.serialization Apache 2.0
  • kotlinx.datetime Apache 2.0

Build & Tooling

  • XcodeGen MIT: iOS project generation
  • Gradle, Build

Full license details per library are available in the app under Settings → Open-Source Libraries.

Back to top

Data Sources

CarbCam uses the following external data sources to validate AI estimates:

USDA FoodData Central

Standard nutritional values for foods

fdc.nal.usda.gov

Open Food Facts

Product database for barcode lookup

openfoodfacts.org

OpenStreetMap / Nominatim

Reverse geocoding for location name (local only)

openstreetmap.org
Back to top

Contact & Feedback

Tip: A clear description plus optionally the meal in question is usually enough for us to handle bugs and support requests.
Back to top