Quantified Self MCP — Claude Code MCP server: install and use
Quantified Self MCP — an MCP server for Claude Code that exposes wearable health logs to AI assistants. Everything runs on local SQLite storage with mathematical validation of observed trends.
MIT ★ 8 updated 2026-11-15
claude mcp add quantified-self-mcp -- uvx quantified-self-mcpWhat Quantified Self MCP does
The server provides structured access to wearable metrics imported from Apple Health, WHOOP, Oura, or Android Health Connect archives. Data stays inside a local SQLite file named health.db, which can be seeded with two months of synthetic metrics using quantified-self-init-db --demo.
A built-in ClaimEvidence system calculates sample density and logging continuity before classifying findings into evidence categories. You can block selected metrics from AI queries by declaring HEALTH_PRIVATE_FIELDS=weight_kg,mood. Readiness of system packages and the database schema is verified with quantified-self-doctor.
What it helps with
- analyze wearable metrics inside a local SQLite database
- validate statistical significance of physiological shifts
- redact sensitive biometric parameters from model prompts
- ingest health exports from Apple Health, WHOOP, and Oura
Example prompts for the agent
- Inspect changes in my resting pulse over the past fortnight
- Determine if recent evening strain influenced night recovery
- Summarize which biometric indicators fluctuated most this quarter
How to install Quantified Self MCP
Platform CLI
claude mcp add quantified-self-mcp -- uvx quantified-self-mcpuvx runs the command named after the package; if the README names another one, use uvx --from quantified-self-mcp <command>
Platform CLI
codex mcp add quantified-self-mcp -- uvx quantified-self-mcpuvx runs the command named after the package; if the README names another one, use uvx --from quantified-self-mcp <command>
Config file
{
"mcpServers": {
"quantified-self-mcp": {
"command": "uvx",
"args": [
"quantified-self-mcp"
]
}
}
}uvx runs the command named after the package; if the README names another one, use uvx --from quantified-self-mcp <command>
Config file
{
"servers": {
"quantified-self-mcp": {
"type": "stdio",
"command": "uvx",
"args": [
"quantified-self-mcp"
]
}
}
}uvx runs the command named after the package; if the README names another one, use uvx --from quantified-self-mcp <command>
Platform CLI
gemini mcp add quantified-self-mcp uvx quantified-self-mcpuvx runs the command named after the package; if the README names another one, use uvx --from quantified-self-mcp <command>
Step by step, with commands for each platform: How to add an MCP server to Claude Code, Codex, Cursor, VS Code, Gemini CLI and more
Questions about Quantified Self MCP
How to install Quantified Self MCP in Claude Code?
Run claude mcp add quantified-self-mcp -- uvx quantified-self-mcp.
What does Quantified Self MCP do?
It lets AI inspect local health records and attaches confidence tiers to detected health changes.
What is required to run Quantified Self MCP?
You need Python with uvx and an initialized SQLite database created via quantified-self-init-db.