ClinicalTrials MCP Server

What is ClinicalTrials MCP Server?

The ClinicalTrials MCP Server bridges AI assistants with ClinicalTrials.gov’s clinical trial repository through the Model Context Protocol (MCP). It allows AI models to search for clinical trials and access their metadata programmatically. Designed for researchers, developers, and health-science professionals who need efficient, code-driven access to trial data.

How to use ClinicalTrials MCP Server?

Install automatically via Smithery (npx -y @smithery/cli@latest install ClinicalTrials-mcp-server --client claude) or manually with uv tool install ClinicalTrials-mcp-server. For development, clone the repo and run uv venv && source .venv/bin/activate && pip install -r requirements.txt. Start the server with python clinical_trials_server.py. Then use the provided MCP tools (e.g., search_clinical_trials_and_save_studies_to_csv) in your AI assistant or application.

Key features of ClinicalTrials MCP Server

• Search trials with custom or advanced search strings.
• Retrieve detailed metadata for specific trials using NCT ID.
• Save, load, and list CSV files of trial data.
• Local storage for faster subsequent access.
• Get study statistics optionally filtered by condition.
• Access trial data via MCP resources (e.g., clinicaltrials://study/{nct_id}).

Use cases of ClinicalTrials MCP Server

• Searching for recent COVID-19 vaccine trials and saving results to CSV.
• Retrieving full details of a specific trial by its NCT ID.
• Filtering studies by keyword (e.g., “diabetes”) and exporting to CSV.
• Getting aggregate statistics on clinical trials for a given condition.
• Loading previously saved CSV data for offline analysis or reporting.

FAQ from ClinicalTrials MCP Server

What dependencies does ClinicalTrials MCP Server require?

Python 3.10+, FastMCP, pytrials, and pandas. Install them with pip install FastMCP pytrials pandas.

How do I install ClinicalTrials MCP Server?

You can install automatically via Smithery (npx ... install ClinicalTrials-mcp-server) or manually with uv tool install ClinicalTrials-mcp-server. See the Quick Start section for details.

Can I use ClinicalTrials MCP Server with Claude Desktop?

Yes. Add the server configuration to your claude_desktop_config.json using the command python -m ClinicalTrials-mcp-server. After restarting Claude, you can ask it to search trials or retrieve details.

What data source does the server use?

All data comes from ClinicalTrials.gov via its API. The tool respects the site’s terms of service and is intended for research purposes only.

Is authentication required to use the server?

No authentication is needed. The server communicates locally via MCP (stdio transport) and queries the public ClinicalTrials.gov API.