What features does FHIR MCP server provide?

Q: What features does FHIR MCP server provide?

MCP-compatible transport: Serves FHIR via stdio, SSE, or streamable HTTP, enabling flexible client connections.. SMART-on-FHIR authentication: Supports SMART-on-FHIR based authentication to securely access FHIR servers using OAuth2 flows and tokens.. Tool integration with MCP clients: Designed to integrate with common MCP clients such as VS Code, Claude Desktop, and MCP Inspector.. OAuth2 configuration for MCP: Configurable via environment variables (e.g., FHIR_SERVER_BASE_URL, FHIR_SERVER_CLIENT_ID, FHIR_SERVER_SCOPES) and optional authorization disablement.. Docker & Docker Compose deployment: Provides a Docker image and docker-compose setup for quick, isolated deployment including a dev stack with HAPI FHIR and PostgreSQL.. FHIR resource toolset: Exposes MCP operations such as get_capabilities, search, read, create, update, delete, and get_user for FHIR resources.. Demo scenarios: Includes demonstrations with HAPI FHIR and Epic Sandbox to illustrate OAuth flows and MCP integration.. MCP Inspector support: Supports visualization and debugging via MCP Inspector to inspect tool invocations and payloads.

FHIR MCP Server 2025: Features & Setup Guide

Overview

The FHIR MCP Server is a Model Context Protocol (MCP) server that provides seamless integration between AI/LLM tools and FHIR healthcare data. It acts as a bridge between modern AI workflows and healthcare data, enabling natural language and programmatic search, retrieval, and analysis across FHIR resources. The server supports multiple transport modalities (streamable HTTP, Server-Sent Events, and STDIO) to accommodate diverse client setups. It uses SMART-on-FHIR based authentication to securely connect to FHIR servers, with configurable OAuth2 parameters and an option to disable authorization for public servers. The MCP server exposes a suite of operations (get_capabilities, search, read, create, update, delete, get_user) against FHIR resource types, enabling tool developers to discover capabilities, run queries, and manipulate resources. It can be run from PyPI or from source, and supports Docker and Docker Compose deployments, including an instant development stack with a HAPI FHIR server and PostgreSQL. It is designed for easy integration with MCP clients such as VS Code, Claude Desktop, and MCP Inspector, facilitating AI-assisted healthcare data workflows.

Features

MCP-compatible transport

Serves FHIR via stdio, SSE, or streamable HTTP, enabling flexible client connections.

SMART-on-FHIR authentication

Supports SMART-on-FHIR based authentication to securely access FHIR servers using OAuth2 flows and tokens.

Tool integration with MCP clients

Designed to integrate with common MCP clients such as VS Code, Claude Desktop, and MCP Inspector.

OAuth2 configuration for MCP

Configurable via environment variables (e.g., FHIR_SERVER_BASE_URL, FHIR_SERVER_CLIENT_ID, FHIR_SERVER_SCOPES) and optional authorization disablement.

Docker & Docker Compose deployment

Provides a Docker image and docker-compose setup for quick, isolated deployment including a dev stack with HAPI FHIR and PostgreSQL.

FHIR resource toolset

Exposes MCP operations such as get_capabilities, search, read, create, update, delete, and get_user for FHIR resources.

Demo scenarios

Includes demonstrations with HAPI FHIR and Epic Sandbox to illustrate OAuth flows and MCP integration.

MCP Inspector support

Supports visualization and debugging via MCP Inspector to inspect tool invocations and payloads.