Anthropic Open Standard

What is MCP?

Model Context Protocol (MCP) is an open standard created by Anthropic to securely connect AI agents to external systems and data.

The Production-Ready Layer

Caren uses MCP to package care pathways, integrations, and governance into a production-ready layer, ensuring seamless interoperability across the clinical ecosystem.

Architecture

Protocol Version 1.0

Architecture Overview

How data flows securely through the Model Context Protocol

AI Agent

Caren Assistant initiating a request for real-time patient context or analysis.

Standardized Bridge

MCP Server

The secure bridge managing context, tool permissions, and protocol transformation.

Clinical Data

Connected systems including EMR, Lab Results, and historical Patient Records.

How it works

Caren provides the abstraction layer between complex clinical logic and modern AI agent infrastructure.

Developer Ready

Universal MCP Standard support

1

Connect data sources

Securely integrate your EHR, data lake, or internal patient database through our encrypted connectors. We handle the semantic mapping to care protocols automatically.

Source Manager Connected
PostgreSQL_Main
FHIR_v4_Endpoint
2

Select or upload a validated care pathway

Choose from our clinical library or upload your institutional protocol files. Our system validates the logic for agent safety and operational constraints.

3

Deploy the MCP server endpoint for agents

One-click deployment transforms pathways into Model Context Protocol (MCP) servers. Your agents can now query, follow, and update tracks with clinical precision.

4

Monitor, audit, and improve outcomes

Real-time monitoring provides full visibility. Track clinical adherence and patient satisfaction through a single pane of glass, then optimize pathways based on real-world data.

Ready to build with MCP?

Explore our documentation and see how Caren integrates MCP to deliver safer, faster AI-driven clinical outcomes.