Our take
Activeregular commits and releases- Activity
- 63
- Community
- 4
- Issues
- 72
- Pull requests
- 68
+3 stars in the last 30 days
- Fewer contributors than 91% of the projects we track
This is an actively maintained, narrowly scoped tool for querying Apple Health CSV exports, with a strong local-data design but development concentrated in one contributor.
Good fit if
- You already use an MCP client and want SQL-backed access to Apple Health data.
- You can export data in the Simple Health Export CSV layout and run Node.js 22 or newer.
- You are comfortable trusting your MCP client with access to your full configured health history.
Look elsewhere if
- Your data is only available as Apple's native export.xml or in another CSV layout.
- You need persistent incremental imports, since the in-memory database is rebuilt each time the server starts.
- You need a project with a broader contributor base, since most of the project work comes from one person.
All health signals
| Last commit | 2026-10-02 (21 days ago) |
|---|---|
| Commits, last 90 days | 83 |
| Releases, last 12 months | 7 (latest v1.4.5, 2026-09-20) |
| Contributors | 1 |
| Issues closed, last 90 days | 10 (typically closed in 1 day) |
| Pull requests merged, last 90 days | 22 (typically merged in 0 days) |
| Project age | 15 months |
Checked on 2026-10-24 with the GitHub API.
Project history
Daily GitHub snapshots since Oct 7, 2026. History covers stars and forks; health signals are not tracked historically.
Stars
577Forks
25Change since Nov 17, 2026: 0 stars·0 forks
Overview
apple-health-mcp is a local-first Apple Health MCP server built on DuckDB. It lets AI assistants access and analyze health data from supported CSV exports, addressing the gap between an Apple Health export and queries an MCP client can use.
It runs locally and reads export files on demand. The project is aimed at people who want to ask questions about sleep, workouts, activity, or heart data through an MCP client, rather than manually inspecting CSV files.
Key Features
- Discovers available tables, columns, units, and sample rows with
health_schema. - Runs a single DuckDB analytical statement with
health_query, returning JSON, CSV, or summary output. - Produces weekly, monthly, or custom summaries with
health_report. - Reads CSV files in place and does not upload the export or make network requests.
- Restricts DuckDB file access to the configured health-data directory and disables network access and temporary disk storage.
- Loads each table's full available history on first use, with a configurable DuckDB memory limit.
Use Cases
- An Apple Health user wants to ask an MCP-connected assistant questions about sleep or activity across their exported history.
- Someone comparing workouts or heart measurements over time wants to use SQL rather than build a separate analysis workflow.
- A quantified-self user wants weekly or monthly summaries from their own health records.
- A developer exploring MCP wants a local server that exposes personal CSV data through schema discovery and analytical tools.
What you need
Detected in the repository
- Node.js >=22.0.0 (from package.json)
- Automated checks on GitHub Actions
License in plain words
MITpermissive
- Commercial use: yes
- Modify and redistribute: yes
- You must keep: the copyright and license notice
- Share your changes: no
A summary, not legal advice: the LICENSE file is what applies.
Getting Started
Configure an MCP client to run the package with npx -y @neiltron/apple-health-mcp and set HEALTH_DATA_DIR to your unzipped export directory. See the README for configuration details and export steps.
Alternatives
- Leo-Health-Core: Parses Apple Health and Whoop exports into SQLite with a terminal view, dashboard, and import watcher, rather than offering SQL queries through an MCP server.
| Project | Language | License | Stars | Status |
|---|---|---|---|---|
| apple-health-mcp | TypeScript | MIT | 576 | Active |
| Leo-Health-Core | Python | MIT | 94 | Quiet |
Considerations
- Only the Simple Health Export CSV layout is supported. The native Apple Health
export.xmlformat is not. - A table's first use loads its full history into memory. The database is rebuilt from CSV files for every server process, and exports that exceed the configured memory limit fail rather than spilling health rows to disk.
- Measurements from overlapping devices can look duplicated, so queries may need to account for
sourceName. - The SQL safeguards reduce accidental side effects but do not isolate the process. The README advises using a trusted local
stdioMCP client and not exposing the server to untrusted network clients. - The project is active, with recent commits and releases and prompt issue and pull-request handling. However, development is concentrated in one contributor, which is a maintenance dependency to weigh for long-term adoption.
Found this useful?
Share it with someone who would like apple-health-mcp.