About the client
The client is a digital health company building a breath-monitoring platform for wellness users. The product depends on signals from the Apple Watch: ECG readings and breath patterns, turned into insights and guided exercises. The client also wanted users to own their data, with a blockchain record of that ownership.
Clean signals from the wrist to the coach, in real time
- Low-latency signal flow. The platform had to capture ECG and breath signals through Apple Watch native APIs and move them with low latency.
- Signal preprocessing. Signals needed accurate preprocessing, filtering and HRV computation both on the device and on the server.
- Store results at scale. The backend had to ingest data continuously, run analysis and store results at scale.
- Coaching and data ownership. On top of that sat breath-pattern detection and coaching logic, an AI chat assistant for guided exercises, a blockchain workflow for user-owned data.
What FiftyFive built
FiftyFive engineered the platform across mobile, backend, AI/ML, blockchain and cloud layers: native watchOS and iOS modules for ECG capture and preprocessing; a Flutter app for a unified cross-platform experience, visualisations and coaching flows; backend services for real-time data pipelines, model inference and analytics; machine learning models for breath-flow classification, HRV insights and personalised recommendations; an AI assistant for interactive coaching and exercise guidance; blockchain workflows that generate NFTs representing users' data ownership; end-to-end encryption, role-based access and secure communication.
How the solution works
Capture on the watch
- Native watchOS and iOS modules read ECG and breath signals through Apple's native APIs and preprocess them on the device, filtering the signal and computing HRV before data leaves the phone.
The app
- The Flutter app receives processed data from the native modules and shows visualisations and coaching flows on one shared interface.
Backend and models
- Backend services ingest data continuously, repeat preprocessing and HRV computation server-side, and run the machine learning models that classify breath flow and generate personalised recommendations.
- An AI assistant guides users through breathing exercises.
Data ownership
- Blockchain workflows mint NFTs that record each user's ownership of their data.
- Communication is secured, data is encrypted end to end and access is role-based.
Technology stack
What the solution enabled
Business transformation
Layers from mobile to cloud
Capture on the Apple Watch
Representing users' data ownership
ECG and breath capture
From the Apple Watch with on-device preprocessing.
A Flutter app
With visualisations and coaching flows.
Classification and HRV insights
Machine learning models for breath-flow classification and HRV insights.
Exercise guidance
An AI coaching assistant for interactive coaching and exercise guidance.
FiftyFive Technologies engineered a breath-monitoring wellness platform for a digital health company that wanted to turn Apple Watch ECG and breath signals into personal coaching. FiftyFive built native watchOS and iOS modules for ECG capture and preprocessing, a Flutter app for visualisation and coaching, backend services for real-time data pipelines and inference, machine learning models for breath-flow classification and heart-rate-variability (HRV) insights, an AI coaching assistant, and blockchain workflows that give users ownership of their data.
Technical questions
How are ECG and breath signals read through the native APIs?
Native watchOS and iOS modules read ECG and breath signals through Apple's native APIs and preprocess them on the device, filtering the signal and computing HRV before data leaves the phone.
How do users own their data in a breath-monitoring app?
Blockchain workflows mint NFTs that record each user's ownership of their data. Communication is secured, data is encrypted end to end and access is role-based.