Withingsmcp
Let a coding agent read your Withings sleep, activity, heart-rate, and weight data locally without building a custom OAuth integration.
Overview
Withings MCP (io.github.davidmosiah/withingsmcp) is an MCP server for the Grow phase that exposes privacy-first Withings health, sleep, activity, and heart-rate data to AI agents over stdio.
What is this MCP server?
- Unofficial stdio MCP server (npm withings-mcp-unofficial) at version 0.4.6 with OAuth via Withings Developer credentials
- Privacy-first design with optional WITHINGS_PRIVACY_MODE and tokens stored under ~/.withings-mcp by default
- Supports setup CLI (withings-mcp-server setup) so client secrets live in ~/.withings-mcp/config.json instead of agent co
- Environment-driven OAuth: WITHINGS_CLIENT_ID, WITHINGS_CLIENT_SECRET, WITHINGS_REDIRECT_URI, WITHINGS_TOKEN_PATH
- Built for AI health, sleep, activity, and heart-rate agent workflows over stdio transport
- Server version 0.4.6
- Single stdio npm package: withings-mcp-unofficial
- Four documented OAuth-related environment variables plus optional WITHINGS_PRIVACY_MODE and WITHINGS_TOKEN_PATH
What problem does it solve?
Agents cannot answer health questions from your wearable data unless you hand-build Withings OAuth, token refresh, and API parsing yourself.
Who is it for?
Indie builders running local MCP agents who want personal or demo health analytics from Withings without a custom integration service.
Skip if: Teams needing official Withings partnership, clinical compliance, or a hosted multi-user health backend.
What do I get? / Deliverables
After OAuth setup, your agent can query Withings metrics from local token storage and summarize trends in natural language.
- Authenticated local Withings token store (~/.withings-mcp/tokens.json by default)
- Agent-callable access to health, sleep, activity, and heart-rate related Withings data
- Optional privacy-mode constrained responses for sensitive metrics
Recommended MCP Servers
Journey fit
Health and habit metrics belong in the Grow phase where builders track signals that inform product and personal performance decisions. Analytics is the right shelf because the server exposes time-series health measurements for summarization and trend questions, not device provisioning or deployment.
How it compares
MCP integration to Withings APIs, not a packaged agent skill or a hosted analytics dashboard.
Common Questions / FAQ
Who is Withings MCP for?
Solo builders and developers who use Claude Code or Cursor and want agents to read their own Withings health data through MCP.
When should I use Withings MCP?
Use it when you are in Grow-phase analytics workflows—sleep, activity, or heart-rate summaries—and prefer local OAuth tokens over a custom API microservice.
How do I add Withings MCP to my agent?
Install the npm package withings-mcp-unofficial, run withings-mcp-server setup or set WITHINGS_CLIENT_ID, WITHINGS_CLIENT_SECRET, and redirect URI, then register the stdio server in your MCP client config.