Medidata MCP Server
Provider-hosted MCP server from Medidata that lets Claude query Medidata platform documentation and run predictive site ranking for clinical trials.
Medidata is a clinical trial software platform (part of Dassault Systèmes) used by life sciences companies to run electronic data capture, study build, data management, and trial intelligence workflows across more than 38,000 historical trials. Medidata hosts an official remote MCP server that exposes a subset of its platform to Claude, allowing trial operations, biostatistics, data management, and study feasibility teams to interact with Medidata data and documentation through natural language.
The server currently provides two capabilities: Platform Help, which answers questions against Medidata's Knowledge Hub documentation for products like Rave EDC, Data Connect, and Clinical Data Studio, and Predictive Site Ranking, which uses Medidata's historical enrollment and site performance data to recommend clinical trial sites for a given indication, phase, and study profile. Site Ranking is restricted to customers of Medidata's Intelligent Trials / Study Feasibility application, while Platform Help is available to any Medidata platform user.
The MCP server is hosted by Medidata at https://mcp.imedidata.com/mcp and is distributed through Anthropic's official anthropics/life-sciences Claude Code plugin marketplace. It is also available as a one-click connector inside Claude for Team and Enterprise customers. Authentication is handled by Medidata's iMedidata identity system, so a user's queries are scoped to the studies and products their Medidata subscription already grants them.
Tools
| Tool | Description |
|---|---|
Platform Help |
Natural language access to Medidata's Knowledge Hub documentation, covering products such as Rave EDC, Data Connect, and Clinical Data Studio. Available to all Medidata platform users. |
Predictive Site Ranking |
Predicts and ranks clinical trial sites against enrollment goals using Medidata's historical trial data, site performance, indication, and phase. Requires a Medidata Intelligent Trials / Study Feasibility subscription. |
Medidata hosts the MCP server, so there is nothing to install or self-host. You connect Claude to https://mcp.imedidata.com/mcp and sign in with your Medidata (iMedidata) account.
Prerequisites
- A Medidata platform account (iMedidata login) for Platform Help.
- A subscription to Medidata Intelligent Trials / Study Feasibility for Predictive Site Ranking.
- Claude Pro, Max, Team, or Enterprise (for the connector UI), or Claude Code (for the plugin install).
Option 1: Add as a Connector in Claude (Team / Enterprise)
For organization owners:
- Go to Admin settings, then Connectors.
- Click "Browse connectors" and select Medidata.
- Choose "Add to your team".
For individual users:
- Go to Settings, then Connectors.
- Find Medidata and click "Connect".
- Complete the iMedidata sign-in flow.
Option 2: Install via Claude Code plugin marketplace
/plugin marketplace add anthropics/life-sciences
/plugin install medidata@life-sciences
Then restart Claude Code and run /mcp to confirm the server is connected. After install, run /plugin, choose "Manage plugins", select Medidata, and click "Configure" to complete authentication.
Option 3: Manual remote MCP configuration
If your MCP client supports remote HTTP MCP servers directly, point it at the Medidata endpoint:
{
"mcpServers": {
"medidata": {
"url": "https://mcp.imedidata.com/mcp"
}
}
}
You will be prompted to authenticate against iMedidata on first use.
- Ask Medidata's documentation in plain English to learn how to configure Rave EDC edit checks, set up Data Connect pipelines, or build dashboards in Clinical Data Studio without leaving Claude.
- During protocol planning, ask Claude to surface the top predicted high performing sites for a given indication and phase based on Medidata's historical enrollment data.
- Stress test a draft protocol by asking which sites rank in the top quartile against more granular study criteria (geography, prior performance, sub-indication).
- Onboard new study managers by letting them query "how do I do X in Medidata?" against the official Knowledge Hub via Claude.
- Speed up RFP and feasibility responses by combining Claude's drafting with Medidata's site intelligence in one conversation.
- "What are the top 10 predicted high performing sites for a Phase III non-small cell lung cancer trial?"
- "Which sites rank in the top 25% for predicted enrollment performance against my study criteria?"
- "How do I create a derived dataset in Data Connect?"
- "Walk me through configuring an edit check in Rave EDC for an out-of-range lab value."
- "Given my draft protocol, which countries and sites should I prioritize for feasibility outreach?"
- Official, vendor-hosted MCP server from Medidata, so there is no third-party scraping or unofficial API use.
- Grounded in Medidata's own clinical trial data set (38,000+ trials, 12M+ patients) and its current product documentation.
- Distributed through Anthropic's curated
anthropics/life-sciencesmarketplace and available as a one-click Claude connector. - Authentication and authorization respect your existing Medidata subscriptions, so users only see data they are already entitled to.
- Only two capabilities are exposed today (Platform Help and Predictive Site Ranking); operational data in Rave EDC, RTSM, eCOA, and similar products is not yet directly queryable.
- Predictive Site Ranking requires a paid Intelligent Trials / Study Feasibility subscription, so it is not usable by non-customers.
- The server is closed source and hosted by Medidata, so it cannot be self-hosted, audited, or extended by customers.
- PubMed MCP server in the same Anthropic life-sciences marketplace for literature search.
- Cortellis Regulatory Intelligence MCP (Clarivate) for regulatory and competitive trial intelligence.
- AdisInsight MCP (Springer Nature) for drug and clinical trial intelligence.