ClinicalTrials

The Clinical Trials MCP Server exposes the ClinicalTrials.gov API v2.0.4 functionality via 18 specialized MCP tools, enabling you to search, filter, and retrieve comprehensive clinical trial data within your MCP-enabled environment. You can perform general and condition-focused searches, fetch detailed study metadata by NCT ID, and explore trials by location, sponsor, intervention, date ranges, and more. Advanced tools allow you to find similar studies, analyze timelines and outcomes, and obtain statistics about trial phases or other groupings. Each tool returns structured JSON data, including study metadata, locations with coordinates, eligibility criteria, contact information, and timeline data, with robust validation and error handling to help you build informed research workflows. To get started, connect the MCP client to the server name you configure (clinical-trials) and invoke any of the supported tools by name with the required parameters.

Typical workflows include: performing a comprehensive search to identify trials matching a condition and geographic area, drilling into a specific NCT ID to retrieve full metadata, obtaining recruiting trials with contact details for outreach, and using the advanced analysis tools to compare studies by similarity, primary outcomes, or eligibility criteria. The API is designed to be developer-friendly, with clear parameter documentation and consistent JSON responses that integrate smoothly into data pipelines and research dashboards.

Prerequisites:

• Node.js 16+ installed on your system
• No API key required (ClinicalTrials.gov API is publicly accessible)

Setup steps:

1. Ensure you have an MCP workspace prepared and cloned from the repository. Navigate to your MCP directory:

cd ~/Documents/Claude/MCP  # or your preferred MCP directory

2. Verify the clinical-trials server directory contents and build artifacts exist after installation/setup (as described in the README):

ls clinical-trials-server/
# Expected: build/ src/ package.json tsconfig.json README.md

3. Install dependencies for the Clinical Trials MCP server (in the server directory if applicable):

cd clinical-trials-server
npm install

4. Build the TypeScript sources to JavaScript (if you are developing from source):

npm run build

5. Run/serve the MCP server (example from the troubleshooting notes):

node build/index.js
# Or, if using a development setup with watch:
# npm run watch

6. Add the MCP server configuration to Claude Desktop configuration (example shown in the README):

{
  "mcpServers": {
    "clinical-trials": {
      "command": "node",
      "args": ["path/to/clinical-trials-server/build/index.js"],
      "autoApprove": [
        "search_studies",
        "get_study_details", 
        "search_by_condition",
        "search_by_location",
        "get_recruiting_studies",
        "search_by_sponsor",
        "search_by_intervention",
        "get_trial_statistics"
      ]
    }
  }
}

Prerequisites recap: ensure Node.js 16+ is installed, and you have unblocked access to the ClinicalTrials.gov API (public).

Environment and configuration tips:

• The server is designed to run under a Claude MCP environment. Ensure the path in the mcp_config matches your deployment path, such as /absolute/path/to/clinical-trials-server/build/index.js.
• If you encounter NCT ID validation errors, verify NCT IDs follow the NCT######## pattern and that your data conforms to the API’s schema.
• Regularly update dependencies and rebuild to stay aligned with API changes from ClinicalTrials.gov.
• The API updates daily Monday through Friday; consider scheduling data refreshes if you cache results locally.
• Use the autoApprove tool list to pre-authorize common queries for convenience, while keeping sensitive operations secured.
• If you run into network timeouts, the server implements timeout protection and retry logic; adjust timeouts if needed for larger result sets.