Skip to main content

CTMS MCP Server

The CTMS MCP Server enables AI assistants to query clinical trial data using the Model Context Protocol. Connect your AI agent to explore cubes, run semantic queries, and get clinical trial KPIs with natural language.

Server Capabilities​

With the MCP server, your agent can:

  • Execute semantic queries against your clinical trial data warehouse
  • Explore cubes, measures, and dimensions
  • Get pre-defined clinical KPIs (enrollment, safety, compliance)
  • Generate SQL for queries without execution

Example prompts:

  • "Show me enrollment counts by study"
  • "How many serious adverse events have been reported?"
  • "Compare subject counts across all sites"
  • "What cubes are available in the semantic layer?"
Enterprise Feature

MCP Server integration is available in Enterprise editions.

Quick Start​

Add to .vscode/mcp.json:

{
  "servers": {
    "ctms-cube": {
      "command": "node",
      "args": ["${workspaceFolder}/ctms-mcp-server/index.js"],
      "env": {
        "CUBE_API_URL": "https://cube.yourdomain.com/cubejs-api/v1"
      }
    }
  }
}

Available Tools​

The MCP server provides the following tools for AI assistants:

📁 list_entities​

List all available data entities in your data warehouse

📁 describe_entity​

Get detailed structure of a specific entity with measures and dimensions

📁 get_schema​

Retrieve full semantic layer metadata including segments and joins

📁 analytics_query​

Execute analytics queries with measures, dimensions, and filters

📁 preview_sql​

Generate SQL for a query without executing it

📁 get_dashboard_kpis​

Get pre-defined dashboard KPIs for reporting

Architecture​

AI Assistant → MCP Server (Node.js) → Cube.dev API → PostgreSQL

For web clients, the Agent Service provides a REST bridge:

Web Client → Agent Service (REST) → MCP Server → Cube.dev