让 Agent 读懂你的身体:数据获取篇
The article outlines seven methods for feeding health data (from Apple Health, Oura, Health Connect, and vendor apps) into AI agents, organized around a three-layer data model (raw metrics, algorithmic scores, manual logs). Practical paths include direct AI app integration (ChatGPT/Claude iOS Apple Health access, limited to US), official connectors (Strava MCP for subscribers, 训记 API Skill with write-back, Hevy Pro API Key, Freddy middleware), and Apple Health exports via Shortcuts (with the Aimee1129/apple-health-shortcuts-mcp template), paid tools (Health Push JSON export, Health Auto Export with REST API/cloud sync), or a custom iOS app using the open_wearables_ios_sdk. The article emphasizes verifying field support, data storage location, and exportability before committing to any connector, and includes reusable AI prompts for configuring Shortcuts and SDK-based export workflows.
- why now
- Strava/训记/Hevy 刚开放 MCP 与 API,接入窗口已到。
- topic
- AI Tools & Agent Workflows
- source
- 少数派