For the past few months we've been building an open-source, self-hosted backend that normalises health and fitness data from multiple wearable providers (Garmin, Whoop, Apple Health, Samsung Health and others) into a single AI-ready REST API - and thought it was worth sharing here too :)

We built this out of our own experience doing custom wearable integrations for clients, and from working with paid SaaS solutions that often lack the flexibility to customise data pipelines or debug what's actually happening under the hood. Surprisingly, there weren't any mature open-source projects filling this gap.

What you get:

- A single normalised API covering 60+ health metrics (cardiovascular, sleep, body composition, activity, respiratory) and 80+ workout types across all supported providers

- A dashboard to manage users, API keys, and sync status — with a full view of each user's health data across connected devices

- An AI layer via MCP, letting you plug users' health data into any LLM or build a custom agent around it

- Open-source mobile apps for SDK-based providers (like Apple Health) so you can quickly test how data flows from device to your platform end-to-end

On the roadmap: more providers, deeper AI features, direct wearable connections via BLE/ANT+ and performance and stability improvements to support enterprise deployments.

Happy to answer any questions. If this solves a problem you've been dealing with - give it a try, open an issue or contribute. All feedback welcome!