Open Wearables

The platform provides a self‑hosted, open‑source stack that ingests data from a range of wearables and health services through a single normalized API. It de‑duplicates and stores the incoming metrics, then applies a set of open algorithms to generate health scores for sleep, recovery, strain, stress, HRV, VO2 max and other indicators. All code is MIT‑licensed and can be run on the user’s own infrastructure, allowing full auditability of the scoring logic.

Built on top of the scoring layer is a health AI engine that performs structured reasoning over the scores. It can detect trends, flag anomalies and connect patterns across multiple metrics, using a modular MCP server that can integrate any large language model without acting as a simple wrapper. The engine produces human‑readable explanations such as “light sleep phase extended, REM below target” and actionable recommendations.

Coaching profiles let developers configure domain‑specific logic for use cases ranging from corporate wellness and performance optimization to clinical monitoring. By reusing the same underlying data and scores, different profiles can apply distinct thresholds and recommendation strategies while maintaining a unified view across all supported devices. This modular, transparent approach is aimed at teams building health‑focused products.