StatPearls MCP Server

A Model Context Protocol (MCP) server that fetches disease information from StatPearls, a trusted source of peer-reviewed medical content.

Give your AI system a relaible source of medical knowledge for its next conversation.

Features

• Searches for diseases and medical conditions on StatPearls
• Retrieve comprehensive, reliable medical information from StatPearls
• Convert HTML content to well-formatted Markdown to make it AI-friendly
• Integrates with AI models via the Model Context Protocol

If you don't already have a Model Context Protocol (MCP) client:

If you are a casual user, you can use Claude Desktop to get started using MCP servers. It is a free and open-source desktop application that allows you to run MCP servers locally and connect to them.

If you are a power user/developer, I recommend using VSCode with the RooCode extension which enables you to connect in MCP servers to your development environment for infinite possibilities!

Installation

Once you have an MCP-capable AI client, you can run this server locally.

The easiest way to get up and running is to download the appropriate executable/binary for your OS from the releases page. This will give you a self-contained executable that you can run without any additional setup.

Place this executable in a directory of your choice. Then simply add the following to your `mcp_settings.json` file:

For Windows:

{
  "mcpServers": {
    ...
    "statpearls": {
      "command": "{path_to_executable_here}\\statpearls-mcp.exe"
    },
    ...
  }
}

#### For Mac/Linux:

{

"mcpServers": {

...

"statpearls": {

"command": "{path_to_executable_here}/statpearls-mcp"

},

...

}

}

### Installing via Smithery

To install statpearls-mcp for Claude Desktop automatically via [Smithery](https://smithery.ai/server/@jpoles1/statpearls-mcp):

npx -y @smithery/cli install @jpoles1/statpearls-mcp --client claude

### For Developers:

You can also run the server from source. This requires [Bun](https://bun.sh/) to be installed on your system.
  1. Clone the repository
  2. Install dependencies (`bun install`)
  3. Compile the server (`bun run build`)
  4. Now you can add the server to your `mcp_settings.json` file:

{

"mcpServers": {

...

"statpearls": {

"command": "node",

"args": [

"{path_to_proj_here}/dist/index.js"

]

},

...

}

}

## Tool Definition

The server provides a single tool:

- **statpearls_disease_info**: Fetches comprehensive, reliable medical information about diseases from StatPearls.

### Input Schema

{

"query": "diabetes",

"format_options": {

"includeToc": true,

"maxLength": 50000

}

}

- `query`: Disease or medical condition to search for (required)
- `format_options`: Optional formatting preferences
  - `includeToc`: Whether to include a table of contents (default: true)
  - `maxLength`: Maximum length of the returned content in characters (default: 50000)

### Example Output

The tool returns formatted Markdown content with:

- Title and source information
- Table of contents (optional)
- Structured sections including etiology, epidemiology, pathophysiology, clinical features, diagnosis, treatment, and prognosis (when available)

## Development

### Project Structure

statpearls-mcp/

├── src/ # Source code

│ ├── index.ts # Main entry point and server setup

│ ├── test-html-parser.ts # Test utility for HTML parser

│ ├── test-statpearls-parser.ts # Test utility for StatPearls parser

│ ├── testrun.ts # Test runner utility

│ ├── tools/ # Tool definitions and handlers

│ │ └── statpearls.ts # StatPearls tool definition and handler

│ ├── services/ # Core functionality services

│ │ ├── search.ts # Search functionality

│ │ ├── content.ts # Content retrieval and processing

│ │ └── markdown.ts # HTML to Markdown conversion

│ ├── types/ # Type definitions

│ │ ├── index.ts # Common type definitions

│ │ └── statpearls.ts # StatPearls-specific type definitions

│ └── utils/ # Utility functions

│ ├── html.ts # HTML parsing utilities

│ ├── error.ts # Error handling utilities

│ └── statpearls-parser.ts # StatPearls content parsing utilities

├── scripts/ # Build and utility scripts

│ ├── build.ts # Build script for creating Node.js compatible bundle

│ ├── compile.ts # Script for compiling executables

│ ├── release.ts # Script for handling releases

│ └── version.ts # Script for managing versioning

├── dist/ # Build output directory (not in repository)

├── package.json # Project configuration and dependencies

├── tsconfig.json # TypeScript configuration

├── bun.lock # Bun dependency lock file

├── README.md # Main project documentation

└── RELEASE-PROCESS.md # Documentation for release process

### Building and Releasing

#### Building

The build process creates a single JavaScript file that can run with vanilla Node.js:

Production build

bun run build

or

bun run build:prod

Development build

bun run build:dev

This creates a bundled file at `dist/index.js` that includes all dependencies.

#### Compiling Executables

You can compile platform-specific executables using Bun's compilation feature:

Compile for all platforms

bun run compile:all

Compile for specific platforms

bun run compile:linux

bun run compile:windows

bun run compile:mac

This creates executable files in the `dist` directory:
- `statpearls-mcp` (default executable)
- `statpearls-mcp-linux-x64` (Linux)
- `statpearls-mcp-windows-x64.exe` (Windows)
- `statpearls-mcp-darwin-x64` (macOS)

#### Releasing

The release process handles versioning, building, compiling, and Git operations:

Release a patch version (bug fixes)

bun run release:patch

Release a minor version (new features, backward compatible)

bun run release:minor

Release a major version (breaking changes)

bun run release:major

This process:
1. Updates the version in package.json
2. Builds the distribution file
3. Compiles executables for all platforms
4. Creates a Git commit with the version number
5. Creates a Git tag for the version
6. Pushes the commit and tag to GitHub

#### Versioning

The project follows semantic versioning. You can check the current version with:

bun run version

## License

This project is licensed under the MIT License - see the LICENSE file for details.