An Android/Kotlin project that reads user-authorized Health Connect data and brings activity, sleep, body, and vital signs into a clearer phone view. Structured workouts add a local system of record and an optional sync path.
Android / Kotlin
Health Connect
Private development
01 / System
Three distinct data paths.
Health readings, structured workouts, and wrist controls follow separate paths. Their boundaries shape what the app can show, store, and sync.
01
Health readings
Permission-scoped Health Connect read
Record mapping and source attribution
On-device dashboard
02
Workout state
Phone workout session
Local Room database and sync outbox
Optional authenticated cloud sync
03
Wear context
Phone-authoritative workout state
Mirrored Wear OS snapshot
Rest control returned to the phone
02 / Health Connect
Origin stays visible.
The app reads the Health Connect record types it has permission to access. Mappers turn those records into dashboard values and retain source metadata for readable attribution. A source label identifies where Health Connect says a record came from; it does not imply a direct integration with that provider.
Read surfaces
Activity, cardio, body measurements, sleep, and vital signs.
Permission boundary
Health Connect access is requested and can be reviewed through the app's rationale flow.
Source boundary
Origin metadata is mapped to readable labels without presenting every source as an app integration.
03 / Persistence
Local work. Optional sync.
The dashboard can read Health Connect on the phone. Structured workout state has a separate local Room database. Cloud persistence is an optional, authenticated path; EDITH Fit does not claim that every record stays on the device.
On device
Workout state in Room
Sessions, exercises, sets, and pending sync changes are stored locally. The phone owns workout state even when the watch is unavailable.
When configured and signed in
Supabase sync
Health Connect history and completed structured workouts have cloud sync paths. Manual and scheduled background work are coordinated around authentication, network state, and retries.
04 / Phone + Wear
The phone owns the workout.
A Wear OS companion mirrors active workout context and sends rest-timer actions back to the phone. The phone remains authoritative for session state and persistence. This is a companion workflow, not standalone watch tracking or automatic exercise detection.
Phone → watch
A versioned workout snapshot keeps current context available at the wrist.
Watch → phone
Rest controls return to the phone and wait for authoritative state before the watch updates its controls.
05 / Validation
Tested at the boundaries.
Mapper, persistence, and sync tests sit alongside recorded physical-device checks.
Data mapping
Tests cover Health Connect mapping, source labels, incremental changes, and failure paths.
Durable state
Room migrations, workout state, cloud outbox behavior, and sync coordination have focused tests.
Real devices
Project validation records cover physical Android and paired Wear hardware. No personal readings or device identifiers are shown here.
06 / Status
Still being built.
Implemented foundations
Permission-scoped Health Connect reading, source-aware dashboard mapping, local structured workouts, optional cloud sync, and the phone-authoritative Wear companion.
Still in development
Cross-source activity consistency, sync integrity, broader device validation, and product polish continue. EDITH Fit is private and has no public download. No medical or clinical validation is claimed.
