shiny-health

Shiny Health Skill

You are an expert in Shiny Health, a .NET MAUI library that provides a unified API for reading and writing health data from Apple HealthKit (iOS) and Android Health Connect.

When to Use This Skill

Invoke this skill when the user wants to:

• Query health metrics (steps, heart rate, calories, distance, weight, height, body fat, blood pressure, oxygen saturation, sleep, hydration)
• Write/log health data (steps, weight, hydration, blood pressure, etc.)
• Observe real-time health data changes (e.g., monitor step counts or heart rate as they are recorded)
• Set up health data access in a .NET MAUI application
• Request health data permissions (read and/or write) on iOS or Android
• Work with time-bucketed health data aggregations
• Understand which health metrics are available cross-platform
• Configure iOS HealthKit entitlements or Android Health Connect permissions

Library Overview

GitHub: https://github.com/shinyorg/health NuGet: Shiny.Health Namespace: Shiny.Health

Shiny Health provides:

• A single IHealthService interface that works on both iOS and Android
• Read and write support for 30+ cross-platform health metrics spanning activity, body, vitals, nutrition, reproductive/cycle tracking, and workouts
• Real-time observation of health data changes via IAsyncEnumerable<HealthResult>
• Time-bucketed aggregate queries at minute, hour, or day intervals
• Permission management with read/write granularity via PermissionType
• AOT-compatible implementation (no .NET reflection)

Setup

1. Install NuGet Package

dotnet add package Shiny.Health

2. Configure in MauiProgram.cs

public static MauiApp CreateMauiApp()
{
    var builder = MauiApp
        .CreateBuilder()
        .UseMauiApp<App>()
        .UseShiny();

    builder.Services.AddHealthIntegration();
    return builder.Build();
}

3. iOS Setup

Your app requires a provisioning profile with HealthKit capabilities enabled.

Info.plist:

<key>UIRequiredDeviceCapabilities</key>
<array>
    <string>healthkit</string>
</array>
<key>NSHealthUpdateUsageDescription</key>
<string>We need access to update your health data</string>
<key>NSHealthShareUsageDescription</key>
<string>We need access to read your health data</string>

Entitlements.plist:

<key>com.apple.developer.healthkit</key>
<true />
<key>com.apple.developer.healthkit.background-delivery</key>
<true />

4. Android Setup (Health Connect)

Android uses Health Connect (the replacement for the deprecated Google Fit API). Health Connect requires Android 9 (API 28) or higher.

AndroidManifest.xml:

<!-- Declare which health data your app reads -->
<uses-permission android:name="android.permission.health.READ_STEPS" />
<uses-permission android:name="android.permission.health.READ_HEART_RATE" />
<uses-permission android:name="android.permission.health.READ_TOTAL_ENERGY_BURNED" />
<uses-permission android:name="android.permission.health.READ_DISTANCE" />
<uses-permission android:name="android.permission.health.READ_WEIGHT" />
<uses-permission android:name="android.permission.health.READ_HEIGHT" />
<uses-permission android:name="android.permission.health.READ_BODY_FAT" />
<uses-permission android:name="android.permission.health.READ_RESTING_HEART_RATE" />
<uses-permission android:name="android.permission.health.READ_BLOOD_PRESSURE" />
<uses-permission android:name="android.permission.health.READ_OXYGEN_SATURATION" />
<uses-permission android:name="android.permission.health.READ_SLEEP" />
<uses-permission android:name="android.permission.health.READ_HYDRATION" />
<uses-permission android:name="android.permission.health.READ_MENSTRUATION" />
<uses-permission android:name="android.permission.ACTIVITY_RECOGNITION" />

<!-- Optional: declare which health data your app writes (only include the types you need) -->
<uses-permission android:name="android.permission.health.WRITE_STEPS" />
<uses-permission android:name="android.permission.health.WRITE_HEART_RATE" />
<uses-permission android:name="android.permission.health.WRITE_TOTAL_ENERGY_BURNED" />
<uses-permission android:name="android.permission.health.WRITE_DISTANCE" />
<uses-permission android:name="android.permission.health.WRITE_WEIGHT" />
<uses-permission android:name="android.permission.health.WRITE_HEIGHT" />
<uses-permission android:name="android.permission.health.WRITE_BODY_FAT" />
<uses-permission android:name="android.permission.health.WRITE_RESTING_HEART_RATE" />
<uses-permission android:name="android.permission.health.WRITE_BLOOD_PRESSURE" />
<uses-permission android:name="android.permission.health.WRITE_OXYGEN_SATURATION" />
<uses-permission android:name="android.permission.health.WRITE_SLEEP" />
<uses-permission android:name="android.permission.health.WRITE_HYDRATION" />
<uses-permission android:name="android.permission.health.WRITE_MENSTRUATION" />

<!-- Allow your app to discover Health Connect -->
<queries>
    <package android:name="com.google.android.apps.healthdata" />
</queries>

Requirements:

• The Health Connect app must be installed on the device
• Minimum SDK version must be set to 28 (Android 9)

API Reference

Core Types

// Permission type for read/write access
[Flags]
public enum PermissionType
{
    Read = 1,
    Write = 2,
    ReadWrite = Read | Write
}

// Time interval for bucketed queries
public enum Interval { Minutes, Hours, Days }

// Available health data types
public enum DataType
{
    // numeric (NumericHealthResult)
    StepCount, HeartRate, Calories, Distance,
    Weight, Height, BodyFatPercentage, RestingHeartRate,
    BloodPressure, OxygenSaturation, SleepDuration, Hydration,
    BloodGlucose, BodyTemperature, BasalBodyTemperature, RespiratoryRate,
    Vo2Max, HeartRateVariability, LeanBodyMass, BasalEnergyBurned,
    ActiveEnergyBurned, FloorsClimbed, WheelchairPushes, Speed, Power,

    // categorical / event-based
    MenstruationFlow, SexualActivity, OvulationTest, CervicalMucus, IntermenstrualBleeding,

    // structured records
    Workout, Nutrition
}

// Menstrual flow level (categorical). None is iOS-only; Android maps it to Unspecified.
public enum MenstrualFlow { Unspecified, None, Light, Medium, Heavy }

// Reproductive / cycle-tracking enums
public enum SexualActivityProtection { Unspecified, Protected, Unprotected }
public enum OvulationTestOutcome { Inconclusive, Positive, High, Negative }
public enum CervicalMucusAppearance { Unspecified, Dry, Sticky, Creamy, Watery, EggWhite }

// Workout activity (subset mapped on both platforms; unmapped -> Other) and meal type
public enum WorkoutType { Other, Running, Walking, Hiking, Cycling, Swimming, Rowing, Elliptical,
    StairClimbing, StrengthTraining, HighIntensityIntervalTraining, Yoga, Pilates, Tennis,
    Basketball, Soccer, Baseball, Golf, Boxing, MartialArts, Dancing }
public enum MealType { Unknown, Breakfast, Lunch, Dinner, Snack }

// Categorical / structured result records
public record SexualActivityResult(DateTimeOffset Start, DateTimeOffset End, SexualActivityProtection Protection) : HealthResult(DataType.SexualActivity, Start, End);
public record OvulationTestResult(DateTimeOffset Start, DateTimeOffset End, OvulationTestOutcome Outcome) : HealthResult(DataType.OvulationTest, Start, End);
public record CervicalMucusResult(DateTimeOffset Start, DateTimeOffset End, CervicalMucusAppearance Appearance) : HealthResult(DataType.CervicalMucus, Start, End);
public record IntermenstrualBleedingResult(DateTimeOffset Start, DateTimeOffset End) : HealthResult(DataType.IntermenstrualBleeding, Start, End);
public record WorkoutResult(DateTimeOffset Start, DateTimeOffset End, WorkoutType Workout, double? TotalEnergyKilocalories = null, double? TotalDistanceMeters = null, string? Title = null) : HealthResult(DataType.Workout, Start, End);
public record NutritionResult(DateTimeOffset Start, DateTimeOffset End, MealType Meal = MealType.Unknown, string? Name = null, double? EnergyKilocalories = null, double? ProteinGrams = null, double? CarbohydratesGrams = null, double? TotalFatGrams = null, double? FiberGrams = null, double? SugarGrams = null, double? SodiumGrams = null, double? CholesterolGrams = null) : HealthResult(DataType.Nutrition, Start, End);

// Result for single-value metrics
public record NumericHealthResult(
    DataType DataType,
    DateTimeOffset Start,
    DateTimeOffset End,
    double Value
) : HealthResult(DataType, Start, End);

// Result for blood pressure (dual-value)
public record BloodPressureResult(
    DateTimeOffset Start,
    DateTimeOffset End,
    double Systolic,
    double Diastolic
) : HealthResult(DataType.BloodPressure, Start, End);

// Result for menstruation flow (categorical, event-based)
// IsCycleStart marks the first day of the cycle (iOS only; always false on Android)
public record MenstruationFlowResult(
    DateTimeOffset Start,
    DateTimeOffset End,
    MenstrualFlow Flow,
    bool IsCycleStart = false
) : HealthResult(DataType.MenstruationFlow, Start, End);

IHealthService Interface

public interface IHealthService
{
    // Observe real-time health data changes (forward-only, yields new samples as recorded)
    // iOS: push-based HKAnchoredObjectQuery; Android: polls Health Connect change tokens
    IAsyncEnumerable<HealthResult> Observe(DataType dataType, TimeSpan? pollingInterval = null, CancellationToken cancelToken = default);

    // Request read permissions (backward compatible)
    Task<IEnumerable<(DataType Type, bool Success)>> RequestPermissions(params DataType[] dataTypes);
    // Request read, write, or both permissions (uniform for all types)
    Task<IEnumerable<(DataType Type, bool Success)>> RequestPermissions(PermissionType permissionType, params DataType[] dataTypes);
    // Request per-metric read/write permissions in a single call
    Task<IEnumerable<(DataType Type, bool Success)>> RequestPermissions(params (PermissionType Permission, DataType Type)[] permissions);

    // Write health data
    Task Write(NumericHealthResult result, CancellationToken cancelToken = default);
    Task Write(BloodPressureResult result, CancellationToken cancelToken = default);
    Task Write(MenstruationFlowResult result, CancellationToken cancelToken = default);

    // Activity metrics
    Task<IList<NumericHealthResult>> GetStepCounts(DateTimeOffset start, DateTimeOffset end, Interval interval, CancellationToken cancelToken = default);
    Task<IList<NumericHealthResult>> GetAverageHeartRate(DateTimeOffset start, DateTimeOffset end, Interval interval, CancellationToken cancelToken = default);
    Task<IList<NumericHealthResult>> GetCalories(DateTimeOffset start, DateTimeOffset end, Interval interval, CancellationToken cancelToken = default);
    Task<IList<NumericHealthResult>> GetDistances(DateTimeOffset start, DateTimeOffset end, Interval interval, CancellationToken cancelToken = default);

    // Body metrics
    Task<IList<NumericHealthResult>> GetWeight(DateTimeOffset start, DateTimeOffset end, Interval interval, CancellationToken cancelToken = default);
    Task<IList<NumericHealthResult>> GetHeight(DateTimeOffset start, DateTimeOffset end, Interval interval, CancellationToken cancelToken = default);
    Task<IList<NumericHealthResult>> GetBodyFatPercentage(DateTimeOffset start, DateTimeOffset end, Interval interval, CancellationToken cancelToken = default);
    Task<IList<NumericHealthResult>> GetRestingHeartRate(DateTimeOffset start, DateTimeOffset end, Interval interval, CancellationToken cancelToken = default);

    // Vitals
    Task<IList<BloodPressureResult>> GetBloodPressure(DateTimeOffset start, DateTimeOffset end, Interval interval, CancellationToken cancelToken = default);
    Task<IList<NumericHealthResult>> GetOxygenSaturation(DateTimeOffset start, DateTimeOffset end, Interval interval, CancellationToken cancelToken = default);

    // Lifestyle
    Task<IList<NumericHealthResult>> GetSleepDuration(DateTimeOffset start, DateTimeOffset end, Interval interval, CancellationToken cancelToken = default);
    Task<IList<NumericHealthResult>> GetHydration(DateTimeOffset start, DateTimeOffset end, Interval interval, CancellationToken cancelToken = default);

    // Menstruation flow is categorical/event-based - no interval bucketing, returns individual records
    Task<IList<MenstruationFlowResult>> GetMenstruationFlow(DateTimeOffset start, DateTimeOffset end, CancellationToken cancelToken = default);
}

Supported Metrics

Metric	Unit	iOS (HealthKit)	Android (Health Connect)
Step Count	count	StepCount	StepsRecord
Heart Rate	bpm	HeartRate	HeartRateRecord
Calories	kcal	ActiveEnergyBurned	TotalCaloriesBurnedRecord
Distance	meters	DistanceWalkingRunning	DistanceRecord
Weight	kg	BodyMass	WeightRecord
Height	meters	Height	HeightRecord
Body Fat %	%	BodyFatPercentage	BodyFatRecord
Resting Heart Rate	bpm	RestingHeartRate	RestingHeartRateRecord
Blood Pressure	mmHg	BloodPressureSystolic/Diastolic	BloodPressureRecord
Oxygen Saturation	%	OxygenSaturation	OxygenSaturationRecord
Sleep Duration	hours	SleepAnalysis (category)	SleepSessionRecord
Hydration	liters	DietaryWater	HydrationRecord
Blood Glucose	mg/dL	BloodGlucose	BloodGlucoseRecord
Body Temperature	°C	BodyTemperature	BodyTemperatureRecord
Basal Body Temperature	°C	BasalBodyTemperature	BasalBodyTemperatureRecord
Respiratory Rate	breaths/min	RespiratoryRate	RespiratoryRateRecord
VO2 Max	mL/kg/min	VO2Max	Vo2MaxRecord
Heart Rate Variability	ms	HeartRateVariabilitySDNN	HeartRateVariabilityRmssdRecord¹
Lean Body Mass	kg	LeanBodyMass	LeanBodyMassRecord
Basal Energy Burned	kcal	BasalEnergyBurned	BasalMetabolicRateRecord
Active Energy Burned	kcal	ActiveEnergyBurned	ActiveCaloriesBurnedRecord
Floors Climbed	count	FlightsClimbed	FloorsClimbedRecord
Wheelchair Pushes	count	PushCount	WheelchairPushesRecord
Speed	m/s	WalkingSpeed²	SpeedRecord
Power	watts	CyclingPower²	PowerRecord
Menstruation Flow	flow level	MenstrualFlow (category)	MenstruationFlowRecord
Sexual Activity	protection enum	SexualActivity (category)	SexualActivityRecord
Ovulation Test	result enum	OvulationTestResult (category)	OvulationTestRecord
Cervical Mucus	appearance enum	CervicalMucusQuality (category)	CervicalMucusRecord
Intermenstrual Bleeding	event	IntermenstrualBleeding (category)	IntermenstrualBleedingRecord
Workout	session	HKWorkout	ExerciseSessionRecord
Nutrition	food/macros	Food correlation (dietary types)	NutritionRecord

¹ HRV caveat: iOS reports SDNN, Health Connect reports RMSSD - both are HRV in milliseconds but computed differently, so the values are not directly comparable across platforms. ² Speed/Power caveat: Health Connect's SpeedRecord/PowerRecord are generic; iOS has no generic equivalents, so Speed maps to walking speed and Power maps to cycling power.

The categorical, event-based, and structured metrics differ from the numeric ones: they use their own result records (SexualActivityResult, OvulationTestResult, CervicalMucusResult, IntermenstrualBleedingResult, WorkoutResult, NutritionResult, plus MenstruationFlowResult), have no Interval bucketing, and are read via dedicated methods (GetSexualActivity, GetOvulationTests, GetCervicalMucus, GetIntermenstrualBleeding, GetWorkouts, GetNutrition). On Android, a WorkoutResult's energy/distance are null on read (Health Connect stores those as separate records from the exercise session).

Menstruation flow is different from the other numeric metrics: it is categorical (a MenstrualFlow level, not a double) and event-based, so it uses MenstruationFlowResult, has no Interval bucketing, and is read via GetMenstruationFlow(start, end). iOS exposes a None level and an IsCycleStart flag (persisted via HealthKit cycle metadata); Health Connect has no None value (mapped to Unspecified) and ignores IsCycleStart.

Usage Examples

Request Permissions and Query Data

IHealthService health; // inject via DI

// Request read permissions for the data types you need
var result = await health.RequestPermissions(
    DataType.StepCount,
    DataType.HeartRate,
    DataType.Calories,
    DataType.Distance
);

// Or request per-metric read/write permissions in a single call
var result2 = await health.RequestPermissions(
    (PermissionType.Read, DataType.StepCount),
    (PermissionType.Read, DataType.HeartRate),
    (PermissionType.Write, DataType.Weight),
    (PermissionType.ReadWrite, DataType.BloodPressure)
);

// Check which permissions were granted
foreach (var (type, success) in result)
{
    if (!success)
        Console.WriteLine($"Permission denied for {type}");
}

// Query data for the last 24 hours, bucketed by day
var end = DateTimeOffset.Now;
var start = end.AddDays(-1);

var steps = (await health.GetStepCounts(start, end, Interval.Days)).Sum(x => x.Value);
var calories = (await health.GetCalories(start, end, Interval.Days)).Sum(x => x.Value);
var distance = (await health.GetDistances(start, end, Interval.Days)).Sum(x => x.Value);
var heartRate = (await health.GetAverageHeartRate(start, end, Interval.Days)).Average(x => x.Value);

Query Body Metrics

var weight = (await health.GetWeight(start, end, Interval.Days)).Average(x => x.Value); // kg
var height = (await health.GetHeight(start, end, Interval.Days)).Average(x => x.Value); // meters
var bodyFat = (await health.GetBodyFatPercentage(start, end, Interval.Days)).Average(x => x.Value); // %
var restingHr = (await health.GetRestingHeartRate(start, end, Interval.Days)).Average(x => x.Value); // bpm

Query Vitals

// Blood pressure returns BloodPressureResult with Systolic and Diastolic
var bp = await health.GetBloodPressure(start, end, Interval.Days);
if (bp.Any())
{
    var avgSystolic = bp.Average(x => x.Systolic);   // mmHg
    var avgDiastolic = bp.Average(x => x.Diastolic);  // mmHg
}

var o2 = (await health.GetOxygenSaturation(start, end, Interval.Days)).Average(x => x.Value); // %

Query Lifestyle

var sleep = (await health.GetSleepDuration(start, end, Interval.Days)).Sum(x => x.Value); // hours
var water = (await health.GetHydration(start, end, Interval.Days)).Sum(x => x.Value); // liters

Hourly Breakdown

// Get hourly step counts for the past week
var weekStart = DateTimeOffset.Now.AddDays(-7);
var weekEnd = DateTimeOffset.Now;

var hourlySteps = await health.GetStepCounts(weekStart, weekEnd, Interval.Hours);
foreach (var bucket in hourlySteps)
{
    Console.WriteLine($"{bucket.Start:g} - {bucket.End:g}: {bucket.Value:N0} steps");
}

ViewModel Pattern (with CommunityToolkit.Mvvm)

public partial class HealthDashboardViewModel(IHealthService health) : ObservableObject
{
    [ObservableProperty]
    double steps;

    [ObservableProperty]
    double calories;

    [RelayCommand]
    async Task LoadDataAsync()
    {
        await health.RequestPermissions(DataType.StepCount, DataType.Calories);

        var start = DateTimeOffset.Now.Date;
        var end = DateTimeOffset.Now;

        Steps = (await health.GetStepCounts(start, end, Interval.Days)).Sum(x => x.Value);
        Calories = (await health.GetCalories(start, end, Interval.Days)).Sum(x => x.Value);
    }
}

Writing Health Data

IHealthService health; // inject via DI

// Request write permissions for the data types you need
await health.RequestPermissions(PermissionType.Write, DataType.Weight, DataType.StepCount, DataType.Hydration);

// Or request both read and write at once
await health.RequestPermissions(PermissionType.ReadWrite, DataType.Weight);

// Or mix read/write per metric in a single call
await health.RequestPermissions(
    (PermissionType.Write, DataType.Weight),
    (PermissionType.Write, DataType.StepCount),
    (PermissionType.ReadWrite, DataType.Hydration)
);

var now = DateTimeOffset.Now;

// Write a weight measurement (point-in-time: Start == End)
await health.Write(new NumericHealthResult(DataType.Weight, now, now, 75.0)); // kg

// Write step counts over a time range
await health.Write(new NumericHealthResult(DataType.StepCount, now.AddMinutes(-30), now, 500));

// Write hydration
await health.Write(new NumericHealthResult(DataType.Hydration, now.AddHours(-1), now, 0.5)); // liters

// Write blood pressure
await health.Write(new BloodPressureResult(now, now, 120.0, 80.0)); // mmHg

// Write sleep session
var sleepStart = now.AddHours(-8);
await health.Write(new NumericHealthResult(DataType.SleepDuration, sleepStart, now, 0)); // Value is ignored, duration derived from Start/End

Menstruation / Period Tracking

IHealthService health; // inject via DI

// Request read+write access for menstruation flow
await health.RequestPermissions(PermissionType.ReadWrite, DataType.MenstruationFlow);

var today = DateTimeOffset.Now;

// Log today's flow. IsCycleStart: true marks the first day of the period (iOS persists this;
// Android ignores it - model the period span separately there if needed).
await health.Write(new MenstruationFlowResult(today, today, MenstrualFlow.Medium, IsCycleStart: true));
await health.Write(new MenstruationFlowResult(today.AddDays(1), today.AddDays(1), MenstrualFlow.Light));

// Read back the cycle's records (categorical + event-based, so NO Interval bucketing)
var records = await health.GetMenstruationFlow(today.AddMonths(-1), today);
foreach (var r in records)
    Console.WriteLine($"{r.Start:d}: {r.Flow}{(r.IsCycleStart ? " (cycle start)" : "")}");

Observing Real-Time Health Data

IHealthService health; // inject via DI

// Request read permission for the data type you want to observe
await health.RequestPermissions(DataType.StepCount);

// Observe step count changes in real time using IAsyncEnumerable
// Use a CancellationTokenSource to stop observation when done
using var cts = new CancellationTokenSource();

await foreach (var result in health.Observe(DataType.StepCount, cancelToken: cts.Token))
{
    // result is a HealthResult — cast to NumericHealthResult for value
    if (result is NumericHealthResult numeric)
        Console.WriteLine($"Steps: {numeric.Value} ({numeric.Start:t} - {numeric.End:t})");
}

// On Android, you can customize the polling interval (default 5s, ignored on iOS)
await foreach (var result in health.Observe(DataType.HeartRate, pollingInterval: TimeSpan.FromSeconds(10), cancelToken: cts.Token))
{
    if (result is NumericHealthResult numeric)
        Console.WriteLine($"Heart rate: {numeric.Value} bpm");
}

Platform Notes

iOS

• Observe uses HKAnchoredObjectQuery for push-based real-time updates (no polling needed)
• HealthKit requires a real device (not simulator) for most data types
• RequestPermissions on iOS does NOT tell you if the user denied access (Apple privacy policy) - it may return true even when denied
• Sleep data uses HKCategoryTypeIdentifier.SleepAnalysis (category type, not quantity type) - the library handles this internally
• Blood pressure requires permissions for both systolic and diastolic types - the library handles this automatically
• Percentage values (body fat, O2 saturation) are returned as 0-100, not 0-1

Android

• Observe uses Health Connect change tokens with polling (default 5s interval, configurable via pollingInterval parameter)
• The Health Connect app must be installed on the device
• Body fat percentage and oxygen saturation use ReadRecords instead of aggregate queries (Health Connect does not provide aggregate metrics for these types)
• Sleep duration uses SleepSessionRecord.SleepDurationTotal aggregate metric, returning hours
• Blood pressure uses BloodPressureRecord.SystolicAvg and DiastolicAvg aggregate metrics
• All Kotlin coroutine interop is handled internally via IContinuation bridge (AOT-safe, no reflection)

Best Practices

1. Always request permissions first - Call RequestPermissions before reading or writing data. Use PermissionType.Write or PermissionType.ReadWrite when writing
2. Use appropriate intervals - Use Interval.Days for summaries, Interval.Hours for detailed breakdowns
3. Handle empty results - Check .Any() before calling .Average() to avoid InvalidOperationException
4. Use CancellationToken - Pass cancellation tokens for long-running queries
5. Sum vs Average - Use .Sum() for cumulative metrics (steps, calories, distance, hydration, sleep) and .Average() for point-in-time metrics (heart rate, weight, height, body fat, O2 sat, resting HR)
6. Blood pressure is special - It returns BloodPressureResult (not NumericHealthResult) with separate Systolic and Diastolic values
7. Menstruation flow is special - It is categorical and event-based: use MenstruationFlowResult/MenstrualFlow, read with GetMenstruationFlow(start, end) (no Interval), and remember None/IsCycleStart are iOS-only
8. Register early - Call AddHealthIntegration() in MauiProgram.cs during app startup

AI Tool Integration (Shiny.Health.Extensions.AI)

The optional Shiny.Health.Extensions.AI package exposes IHealthService as Microsoft.Extensions.AI tool functions (AIFunctions) for LLM agents. It uses a small set of parameterized tools (one read tool covers all numeric metrics via a metric enum arg, not one tool per metric). Read-only by default; write is opt-in per area. AOT-compatible (hand-built schemas, JsonNode results — no reflection).

using Shiny.Health;
using Shiny.Health.Extensions.AI;

builder.Services.AddHealthIntegration();          // registers IHealthService
builder.Services.AddHealthAITools(tools => tools
    .AddAllMetrics()                                          // read all numeric metrics
    .AddMetric(DataType.Weight, HealthAICapabilities.ReadWrite)
    .AddBloodPressure(HealthAICapabilities.ReadWrite)
    .AddCycleTracking()                                       // read cycle records
    .AddWorkouts(HealthAICapabilities.ReadWrite)
    .AddNutrition()
);

// resolve the bundle and pass the tools to any IChatClient
var tools = sp.GetRequiredService<HealthAITools>().Tools;
var response = await chatClient.GetResponseAsync(
    messages,
    new ChatOptions { Tools = [.. tools] }
);

Key types:

• AddHealthAITools(Action<IHealthAIToolBuilder>) — DI extension; throws if nothing is added.
• IHealthAIToolBuilder — AddMetric(DataType, capabilities), AddAllMetrics(...), AddBloodPressure(...), AddCycleTracking(...), AddWorkouts(...), AddNutrition(...). AddMetric throws for non-numeric DataTypes.
• HealthAICapabilities [Flags] — None, Read (default), Write, ReadWrite.
• HealthAITools — resolve from DI; .Tools is IReadOnlyList<AITool>.

Generated tools (only for opted-in areas; enum args constrained to what you allowed): get_health_metric, write_health_metric, get_blood_pressure, write_blood_pressure, get_cycle_records (kind enum), write_menstruation_flow, get_workouts, write_workout, get_nutrition, write_nutrition. Dates are ISO-8601; interval is minutes/hours/days.

The AI tools assume permissions are already granted — they do not trigger the platform permission UI (needs a foreground activity). Call IHealthService.RequestPermissions(...) from the app before invoking the agent.

Common Packages

dotnet add package Shiny.Health                 # Core health data library
dotnet add package Shiny.Health.Extensions.AI   # Optional: Microsoft.Extensions.AI tool surface for LLM agents
dotnet add package Shiny.Core                   # Required dependency