$ show StanfordBDHG/HealthGPT
HealthGPT is an iOS application that enables users to interact with their Apple Health data using large language models (LLMs) in a conversational interface. Built on the Stanford Spezi modular framework, the app supports three LLM execution modes: remote OpenAI (GPT-3.5/4), on-device local models (Llama3 via MLX), and privacy-preserving fog nodes within the local network. The architecture follows a monolith application pattern with clear separation of concerns: HealthDataFetcher queries HealthKit for steps, sleep, heart rate, etc.; HealthDataInterpreter prepares the LLM context with a system prompt generated from the fetched data; and the SwiftUI ChatView manages the conversation flow. External dependencies include Spezi modules (HealthKit, LLM, Onboarding, Storage) and OpenAI API. The codebase is well-structured with SwiftLint enforcement, comprehensive UI and unit tests, CI/CD pipelines via GitHub Actions and Fastlane, and an automated TestFlight deployment workflow. While the app is tightly coupled to the iOS ecosystem and Apple HealthKit, the data fetching, prompt generation, and LLM orchestration patterns are cleanly abstracted and potentially reusable in other health-related Swift applications.
The LLM identified 2 files worth highlighting for reuse:
Encapsulates HealthKit queries for the last two weeks of step count, active energy, exercise minutes, body weight, resting heart rate, and sleep analysis. Uses the modern Swift async/await API (HKStatisticsCollectionQueryDescriptor, HKSampleQueryDescriptor) to fetch data efficiently. The class is an Observable module registered in the Spezi dependency injection system, making it injectable into any Spezi-based app. The generic fetchLastTwoWeeksQuantityData method is parameterized by identifier, unit, and options, allowing extension to any HKQuantityType. The sleep query logic incorporates a custom time window (3PM to 3PM) for accurate daily sleep tracking. This is a clean, self-contained component that could be extracted with minimal changes for any iOS app needing HealthKit data aggregation.
Constructs a structured system prompt for the LLM from daily health data. It takes an array of HealthData objects (each representing one day) and builds a human-readable text that includes step counts, sleep hours, active energy, exercise minutes, body weight, and resting heart rate per day. The main prompt instructs the LLM to act as a caretaker, avoid statistics, and provide advice when numbers are low. This is a self-contained class with no internal dependencies (only Foundation and the HealthData model). It demonstrates a clear pattern of converting structured data into a natural language prompt that can be reused in any LLM-based health application, regardless of the data source or LLM provider.