clintrials-mcp

An MCP (Model Context Protocol) server that provides structured access to ClinicalTrials.gov data, allowing LLMs to search, retrieve, and analyze clinical trial information.

• Multi-parameter Search: Search trials by condition, intervention, sponsor, NCT ID, or combination
• Detailed Retrieval: Get comprehensive trial details including results, eligibility, outcomes
• Statistical Analysis: Analyze trial phases, field statistics, and data distributions
• Flexible Field Selection: Request specific data fields to optimize responses
• Error Handling: Robust error handling with meaningful error messages

1. Install dependencies (Python runtime is required):

pip install -r requirements.txt

> Note: JavaScript package managers like npm, pnpm, or bun are not required for this project because all runtime dependencies are Python packages managed through requirements.txt.

1. Run the MCP server:

python mcp_server.py

Cost: FREE for first 1M requests/month, then $0.20 per 1M requests Setup time: 2 minutes

1. Install AWS CLI and configure credentials:

   aws configure
   # Enter your AWS Access Key ID, Secret Key, and region

1. Deploy to Lambda:

   ./deploy-lambda.sh
   # Takes ~60 seconds, automatically creates function + public URL

1. Copy the Function URL and configure Cloudflare:

   npx wrangler secret put BACKEND_URL
   # Paste your Lambda Function URL when prompted

1. Deploy Cloudflare Worker:

   npx wrangler deploy

Your MCP server will be available at: https://<worker-name>.workers.dev/sse

Pros: Completely free tier, auto-scales, no server management Cons: Cold starts (~1-2s first request after idle)

*