saksham2001/PulseLoopiOS

PulseLoop – AI‑enhanced health tracker for iOS

What it is – PulseLoop is an open‑source iOS app (Swift / SwiftUI) that talks directly to inexpensive Bluetooth “smart rings” (e.g. generic J‑ring, Colmi/Yawell R02/R0x/R1x/H59, R10M, TK5). It reads raw sensor data – heart‑rate, SpO₂, steps, distance, calories, sleep stages – without using any vendor cloud. The app stores the data locally with SwiftData and displays native dashboards.

AI Coach – On top of the raw data PulseLoop can run an LLM‑based coach. The coach can be any OpenAI‑compatible service (OpenAI, Gemini, OpenRouter, self‑hosted Ollama, llama.cpp, vLLM, etc.) and is equipped with tools for:

• retrieving the user’s health metrics,
• performing on‑the‑fly analysis and chart generation,
• remembering long‑term context, and
• web‑search or image‑input (e.g., a photo of a meal). All responses are grounded in the user’s actual ring data, turning static charts into a conversational health assistant.

Key features

• BLE driver layer that decodes each ring’s proprietary protocol.
• SwiftUI dashboards: Today, Vitals, Sleep, Activity, Workout summary.
• Live workout mode with heart‑rate zones, GPS routes, Dynamic Island widget.
• Daily check‑in notifications generated by the AI Coach.
• Cross‑platform support – a sister Android repo shares the BLE code; iOS leads on the AI Coach.

Who it’s for – Anyone who wants to replace a pricey subscription‑based wearable with a $20‑$30 ring plus on‑device intelligence, and who prefers the AI logic to run locally or on a self‑hosted LLM rather than a vendor cloud.

Getting started

• TestFlight: public beta available via a sign‑up form.
• Build from source: Xcode 16+, iOS 18+, a compatible ring, and optionally an OpenAI/Gemini API key for the coach. Detailed docs walk through signing, bundle ID setup, and demo data.

Community – Discord server for roadmap, support, and coordination between iOS and Android contributors. Good first issues are labelled in the repo.

License – Creative Commons Attribution 4.0 (CC‑BY‑4.0).

---

PulseLoop demonstrates a concrete use‑case of LLMs as personal health assistants, keeping the intelligence on the phone and leveraging cheap hardware.