One MCP server, eight AI clients: Apple Health setup that works everywhere

A developer has released an MCP server for Apple Health that works across eight AI clients, including Claude, Cursor, and ChatGPT, using a single server command. The server provides eight read-only tools for querying 190 health metrics locally, with no network calls, and the configuration only varies by file path and JSON shape per client.

Most Apple Health MCP guides are per-client. Claude today, Cursor tomorrow, ChatGPT next week. That's the wrong abstraction. MCP is a protocol. The server command is identical across every client. Only the config file path and JSON shape change. Eight read-only tools, all local: | Tool | What it returns | |---|---| get mcp status | Health check: data source, metric count, most recent data date | list metrics | Every available metric with unit, day count, date range | get health metrics | Values for a metric over a date range with aggregation | get trends | Compare recent N-day window against prior N | compare periods | A/B two arbitrary date windows for a single metric | get structured export | Clean structured JSON, paginated | get intraday | Hour-by-hour window, live within-day view | query health data | Natural-language convenience query | 190 Apple Health metrics. 84 kB npm package. Zero dependencies. Zero network calls. Claude Desktop ~/Library/Application Support/Claude/claude desktop config.json : { "mcpServers": { "health-export": { "command": "node", "args": "/Users/you/.health-export-mcp/server.mjs" , "env": { "HEALTH DATA DIR": "/Users/you/Library/Mobile Documents/iCloud~ai~healthexport~app/Documents" } } } } Cursor uses the same shape at ~/.cursor/mcp.json . Windsurf uses it at ~/.codeium/windsurf/mcp config.json . ChatGPT Desktop takes the inner block in Settings MCP Servers. Codex CLI reads .mcp.json . opencode uses mcp + environment + array command instead of mcpServers + env . The server command never changes. Only the file path does. You: What is my HRV trend over the last 14 days? Agent: calls get trends Your HRV averaged 58 ms, up 5.5% vs prior 14 days. You: Compare my sleep this week to last week. Agent: calls compare periods Deep sleep up 17%, total sleep up 6%. You: How does my RHR change on days I sleep less than 6 hours? Agent: calls query health data RHR averages 58 bpm after <6h sleep, 53 bpm after >7h. No CSV, no copy-paste, no screenshots. Live data from your iPhone, answered locally. Full config blocks for all eight clients, troubleshooting, and FAQ: complete guide on healthexport.dev https://www.healthexport.dev/blog/query-apple-health-with-any-ai-agent?utm source=devto&utm medium=social&utm campaign=exp-20260818-hea80-multiclient