Langcare Mcp Fhir

LangCare MCP FHIR Server

[GitHub Stars] [Contributors] [License] [Go Version]

Enterprise-grade MCP Server for FHIR-based EMRs. Fully written in Go with enterprise-grade security and 4 generic FHIR operations that work with any FHIR R4 resource type. Supports EPIC, Cerner, OpenEMR, GCP Healthcare API, and any generic FHIR R4 server.

Ships with a 40+ Clinical Skills Library — agent-agnostic workflow guides covering medication management, lab interpretation, clinical decision support, documentation, population health, and more. Extend with ✨ New: Claude Managed Agents (9 production-ready clinical AI agents on the Anthropic Managed Agents API), MCP Apps (interactive clinical UIs inside Claude Desktop), a Healthcare Voice Agent (real-time voice AI over FHIR), and a LangCare CLI for agent frameworks that don't speak MCP natively.

[LangCare MCP FHIR — Healthcare AI Agents with FHIR MCP Server for EMRs]
langcare.ai

Installation

Install via npm:

npm install -g @langcare/langcare-mcp-fhir

Or use directly without installation:

npx @langcare/langcare-mcp-fhir -config /path/to/config.yaml

Quick Configuration

LangCare MCP FHIR connects Claude to your FHIR-based EMR system. You need a YAML configuration file pointing to your backend.

1. Get a Config Template

Choose your backend:

• EPIC: config.epic.example.yaml
• Cerner: config.cerner.example.yaml
• OpenEMR: config.openemr.example.yaml
• GCP Healthcare API: config.gcp.example.yaml
• Any FHIR R4 Server: config.base.example.yaml

2. Configure Claude Desktop

Add to your Claude Desktop config file (~/.config/Claude/claude_desktop_config.json):

{  "mcpServers": {    "langcare-mcp-fhir": {      "command": "langcare-mcp-fhir",      "args": ["-config", "/path/to/your/config.yaml"]    }  }}

On macOS, the config is typically at:

~/Library/Application\ Support/Claude/claude_desktop_config.json

3. Restart Claude Desktop

Close and reopen Claude Desktop. The FHIR tools will now be available.

Need detailed setup help? See the Local Testing Guide.

Architecture

This MCP server acts as an intelligent proxy between AI agents and FHIR R4 servers. It exposes 4 generic FHIR operations through the Model Context Protocol (MCP), enabling AI-powered workflows for any FHIR resource type.

Key Design:

• MCP SDK: Official github.com/modelcontextprotocol/go-sdk (Anthropic/Google maintained)
• FHIR Client: Generic HTTP client working with any FHIR R4 server
• Transport: stdio and Streamable HTTP
• Backend: Proxy to existing FHIR server (no database)
• Language: 100% Go for high performance and reliability

4 Generic MCP Tools

All tools work with any FHIR resource type (Patient, Observation, Medication, etc.):

1. fhir_read

Read a FHIR resource by type and ID.

{  "resourceType": "Patient",  "id": "example-123"}

2. fhir_search

Search FHIR resources with query parameters.

{  "resourceType": "Patient",  "queryParams": "name=John&birthdate=gt1990-01-01"}

3. fhir_create

Create a new FHIR resource.

{  "resourceType": "Observation",  "resource": {    "resourceType": "Observation",    "status": "final",    "code": { ... },    "subject": { "reference": "Patient/123" }  }}

4. fhir_update

Update an existing FHIR resource.

{  "resourceType": "Patient",  "id": "example-123",  "resource": {    "resourceType": "Patient",    "id": "example-123",    "name": [{ "family": "Smith" }]  }}

Security Architecture

LangCare MCP FHIR implements a two-layer security model for HIPAA-compliant healthcare data access:

┌─────────────┐         ┌──────────────┐         ┌─────────────┐│   Claude    │ Auth1   │  MCP Server  │ Auth2   │  FHIR API   ││   Client    │────────▶│   (Go)       │────────▶│   (EMR)     │└─────────────┘         └──────────────┘         └─────────────┘
Auth1: MCP Client Authentication (Bearer Token/API Key)Auth2: FHIR Backend Authentication (Bearer/OAuth2/SMART on FHIR)

Security Features

• ✅ TLS 1.3 encryption for HTTP transport
• ✅ PHI Scrubbing in logs (enabled by default)
• ✅ HIPAA-compliant audit logging
• ✅ No persistent PHI storage (stateless proxy)
• ✅ Secrets via environment variables (never in config files)
• ✅ OAuth 2.0 with automatic token refresh
• ✅ mTLS support for service-to-service communication
• ✅ Rate limiting per client

Supported Authentication Methods

• Bearer Token - Simple API key authentication
• OAuth2 - Full OAuth2 flow with token refresh
• SMART on FHIR Backend Services - private_key_jwt (RS384) for EPIC, OpenEMR, and other SMART-conformant EMRs
• SMART on FHIR - EPIC, Cerner, OpenEMR, and other EMR standards
• Basic Auth - Username/password authentication
• Custom - Extensible for additional auth methods

For complete security documentation, see Security Guide:

• HIPAA compliance checklist
• OAuth configuration for EPIC/Cerner/GCP
• Kubernetes security manifests
• Credential management procedures
• Audit logging implementation

MCP Apps (Interactive UIs)

LangCare MCP FHIR ships with built-in MCP Apps — interactive, rich UI views that run directly inside MCP-capable hosts like Claude Desktop. Unlike traditional chat-based tool output, MCP Apps render full React-based interfaces with charts, tables, and interactive controls while using the same underlying FHIR tools.

How it works: Each app is a single-file HTML bundle (React + TypeScript, compiled with Vite) that gets embedded into the Go binary at compile time via go:embed. At runtime the MCP server registers each app as both an MCP Resource (text/html;profile=mcp-app) and a dedicated MCP Tool linked via _meta.ui.resourceUri. When an MCP host calls the tool, it fetches the resource and renders the UI. The app calls back into the server's generic FHIR tools (fhir_search, fhir_read, etc.) through app.callServerTool() — no LLM round-trips for data fetching.

Advantages over plain tool output:

• Rich visualization — SVG charts, color-coded cards, expandable detail panels
• Interactive controls — search fields, date range pickers, click-to-expand rows
• Deterministic data fetching — apps call FHIR tools directly, no LLM involvement in data retrieval
• Zero external dependencies — everything inlines into a single HTML file, embedded in the binary
• Works offline — no CDN, no external scripts, no network requests beyond FHIR API calls

Built-in Apps

Both apps are reference implementations demonstrating the MCP Apps pattern. See apps/README.md for architecture details and how to build new apps.

[Patient Chart Review — clinical dashboard with vitals trends, conditions, medications, and labs]
Patient Chart Review running inside Claude Desktop

Agent Usage

AI agents use LangCare MCP FHIR Server to help healthcare professionals access and manage patient health records through 4 FHIR tools. The server handles EMR authentication, allowing agents to focus on clinical workflows while maintaining strict privacy and accuracy standards.

Agent capabilities:

• Search, Read, Create, Update - Any FHIR R4 resource (Patient, Observation, Medication, etc.)
• Patient privacy - Use partial identifiers, confirm identity before updates
• Clinical accuracy - Verify data, use standard codes (LOINC, SNOMED, RxNorm)
• Professional communication - Structure responses with context, findings, and next steps

Common workflows:

• Patient lookup: Search by name/DOB → verify identity → read full details
• Clinical review: Retrieve labs, vitals, medications → present with reference ranges
• Documentation: Extract structured data → map to FHIR resources → confirm → create
• Updates: Verify existing resource → modify → confirm changes → update

System support:

• Works with any FHIR R4 resource type (60+ types including DocumentReference, Binary, Media)
• Automatic authentication and token refresh to EPIC, Cerner, OpenEMR, GCP Healthcare API
• HIPAA-compliant PHI handling with audit logging
• Comprehensive OAuth2 scopes for clinical data access

📖 Complete guide: Agent Prompt Guide - System prompt, tool examples, workflows, and error handling

Clinical Skills Library (Optional)

40+ agent-agnostic clinical workflow guides that teach AI agents how to perform complex healthcare tasks using the MCP server's 4 FHIR tools (fhir_search, fhir_read, fhir_create, fhir_update).

• Optional - The MCP server works without them
• Portable - Work with Claude, ChatGPT, Gemini, or any AI agent
• Evidence-based - Built on USPSTF, ADA, ACC/AHA, CDC, ACOG, KDIGO, and other society guidelines
• Copy-paste ready - Add a skill's SKILL.md to your agent's system prompt or custom instructions

Skill Categories (40 Skills)

Full catalog with links: skills/README.md

How to Use Skills

1. Browse the skills/core/ directory and pick a skill
2. Copy the skill's SKILL.md content into your AI agent's system prompt or custom instructions
3. Reference files in each skill's references/ subdirectory contain detailed clinical knowledge (scoring criteria, code tables, thresholds) that can optionally be included for deeper clinical accuracy

# Example: Add medication-reconciliation skill to your agentskills/core/medication-management/medication-reconciliation/├── SKILL.md              # Copy this into agent instructions└── references/    ├── reconciliation-process.md   # Joint Commission standards    └── high-risk-medications.md    # ISMP high-alert drug list

Integration guides: Claude | ChatGPT | Gemini

Community contributions welcome - see CONTRIBUTING.md for guidelines.

Development & Testing

Build from Source

make build

Run Locally (stdio mode)

make run# or./bin/langcare-mcp-fhir -config configs/config.local.yaml

Run in HTTP Mode (Streamable HTTP)

make run-http# or./bin/langcare-mcp-fhir -http -port 8080 -config configs/config.yaml

Starts the server with Streamable HTTP transport on /mcp and health check on /health.

Run Tests

make test

Lint Code

make lint

Deploy to Fly.io (Remote Streamable HTTP)

Deploy as a remote MCP server with Streamable HTTP transport, accessible by any MCP-compatible AI agent from anywhere.

# Install Fly CLIbrew install flyctlfly auth login
# Create appfly apps create --name langcare-mcp-dev
# Set CONFIG_FILE in fly/fly.dev.toml [env] block for your provider (EPIC or GCP)# Then set secrets (EPIC example):fly secrets set \  EPIC_BASE_URL="https://fhir.epic.com/interconnect-fhir-oauth/api/FHIR/R4" \  EPIC_CLIENT_ID="your-client-id" \  EPIC_TOKEN_URL="https://fhir.epic.com/interconnect-fhir-oauth/oauth2/token" \  EPIC_PRIVATE_KEY_B64="$(base64 < keys/epic/private-key.pem)" \  MCP_AUTH_TOKENS="your-token" \  --app langcare-mcp-dev
# Deployfly deploy -c fly/fly.dev.toml --app langcare-mcp-dev
# Verifycurl https://langcare-mcp-dev.fly.dev/health

Connect any MCP client to:

URL:   https://langcare-mcp-dev.fly.dev/mcpAuth:  Authorization: Bearer your-token

Claude Desktop (claude_desktop_config.json):

{  "mcpServers": {    "langcare-fhir": {      "url": "https://langcare-mcp-dev.fly.dev/mcp",      "headers": {        "Authorization": "Bearer your-token"      }    }  }}

Supports EPIC and GCP Healthcare API providers. See fly/README.md for provider setup, secrets, and full deployment guide.

Local Testing with EPIC

For step-by-step instructions on setting up EPIC credentials and testing locally:

📖 Local Testing Guide

This guide covers:

• Generating RSA keys and JWKS
• Configuring EPIC credentials
• Running the server locally
• Testing with Claude Desktop
• Troubleshooting common issues

Quick credential test:

# Test your EPIC credentials before running the servergo run test/test_epic_token.go "your-client-id" "/path/to/private-key.pem"

Project Structure

langcare-mcp-fhir/├── cmd/│   └── server/│       └── main.go                          # Entry point├── internal/│   ├── apps/                                # MCP Apps (embedded UIs)│   │   ├── embed.go                         # go:embed directive for HTML bundles│   │   ├── registry.go                      # App metadata, tool names, resource URIs│   │   └── dist/                            # Built HTML bundles (copied by build)│   │       ├── fhir-explorer.html           # FHIR Explorer single-file bundle│   │       └── patient-chart-review.html    # Patient Chart Review single-file bundle│   ├── audit/│   │   └── logger.go                        # HIPAA audit logging│   ├── config/│   │   └── config.go                        # YAML configuration loading│   ├── fhir/│   │   ├── client.go                        # FHIR HTTP client interface│   │   ├── types.go                         # FHIR client types│   │   └── providers/                       # Backend implementations│   │       ├── base.go                      # Base HTTP provider│   │       ├── epic.go                      # EPIC OAuth2 provider│   │       ├── cerner.go                    # Cerner OAuth2 provider│   │       ├── openemr.go                   # OpenEMR SMART Backend Services provider│   │       └── gcp.go                       # GCP Healthcare API provider│   ├── mcp/│   │   └── server.go                        # MCP server + app registration│   ├── middleware/│   │   ├── auth.go                          # MCP authentication│   │   └── rate_limit.go                    # Rate limiting│   ├── tools/                               # MCP tool implementations│   │   ├── registry.go                      # Tool registry│   │   ├── fhir_read.go                     # Read FHIR resource│   │   ├── fhir_search.go                   # Search FHIR resources│   │   ├── fhir_create.go                   # Create FHIR resource│   │   └── fhir_update.go                   # Update FHIR resource│   └── transport/│       ├── stdio.go                         # stdio transport (Claude Desktop)│       └── http.go                          # Streamable HTTP transport (production)├── apps/                                    # MCP App source code (React + TypeScript)│   ├── README.md                            # App development guide│   ├── package.json                         # Shared dependencies (React 19, MCP Apps SDK)│   ├── vite.config.ts                       # Vite build config (single-file output)│   ├── tsconfig.json                        # TypeScript config│   ├── fhir-explorer/                       # FHIR Explorer app│   │   ├── index.html│   │   └── src/│   │       ├── app.tsx│   │       └── global.css│   └── patient-chart-review/                # Patient Chart Review app│       ├── index.html│       └── src/│           ├── app.tsx│           └── global.css├── scripts/│   ├── build-apps.sh                        # Build all apps → internal/apps/dist/│   ├── create_jwks.sh                       # Generate JWKS from public key (EPIC)│   └── create_jwks_openemr.sh               # Generate JWKS from public key (OpenEMR)├── pkg/│   └── types/│       └── errors.go                        # Custom error types├── configs/│   ├── config.epic.example.yaml             # Example configuration for EPIC│   ├── config.cerner.example.yaml           # Example configuration for Cerner│   ├── config.openemr.example.yaml          # Example configuration for OpenEMR│   ├── config.gcp.example.yaml              # Example configuration for GCP│   └── config.base.example.yaml             # Example configuration for any FHIR R4 server├── docs/│   ├── AGENT_PROMPT.md                      # AI agent system prompt│   ├── EPIC-APP-SECURITY.md                 # EPIC authentication setup│   ├── OPENEMR-APP-SECURITY.md              # OpenEMR SMART Backend Services setup│   ├── EPIC-SCOPES.md                       # OAuth2 scopes reference│   ├── LOCAL-TESTING.md                     # Local development guide│   └── SECURITY.md                          # Production security guide├── test/│   ├── README.md                            # Test documentation│   └── test_epic_token.go                   # EPIC OAuth2 token tester├── fly/│   ├── Dockerfile                           # Multi-stage Go build for Fly.io│   ├── docker-entrypoint.sh                 # Key materialization + server startup│   ├── fly.dev.toml                         # Fly.io dev deployment config│   ├── config.fly.epic.yaml                 # Fly.io EPIC provider config│   ├── config.fly.gcp.yaml                  # Fly.io GCP provider config│   └── README.md                            # Fly.io deployment guide├── bin/                                     # Build output (gitignored)│   └── langcare-mcp-fhir                    # Compiled binary├── go.mod                                   # Go module definition├── go.sum                                   # Go module checksums├── Makefile                                 # Build commands└── README.md                                # This file

Note: The following are gitignored and not committed:

• keys/ - Private keys and credentials
• config.local.*.yaml - Local configuration files
• bin/ - Compiled binaries
• .env - Environment variables
• apps/node_modules/, apps/dist/, apps/dist-tmp/ - App build artifacts

Healthcare Voice Agent

Real-time voice AI that lets patients ask about their health records and get spoken answers pulled directly from their EMR.

[Healthcare Voice Agent Architecture — PipeCat + LangCare MCP]

The stack: PipeCat (open-source, Daily.co) for the voice pipeline — STT, LLM orchestration, TTS with sub-3-second latency. Claude for clinical reasoning and tool calling. LangCare MCP FHIR Server (open-source, Go) as a stateless proxy to any FHIR R4 EMR — Epic, Cerner, GCP Healthcare API.

MCP is the glue. PipeCat's native MCP client auto-discovers FHIR tools at startup. Patient asks "What medications am I on?" — Claude calls fhir_search — PipeCat routes it to the MCP server — data comes back — Claude responds in natural speech. No manual tool schemas needed.

Three-layer HIPAA auth: Caller identity verified before the session starts, bearer token to MCP, OAuth2/SMART on FHIR to EMR. Zero PHI storage.

Everything is swappable. Replace Claude with Gemini, DeepGram with Google STT, Daily with WebSocket. The MCP FHIR layer and clinical prompts stay the same.

Full documentation and setup guide

LangCare CLI

Command-line interface that wraps the 4 FHIR MCP tools (fhir_search, fhir_read, fhir_create, fhir_update) as CLI subcommands over HTTP. Built for AI agent frameworks that don't speak MCP natively — LangChain, smolagents, CrewAI, AutoGen, and any framework that can call a subprocess. The CLI handles the MCP session handshake internally, so agents get clean JSON on stdout with no protocol knowledge required.

[LangCare CLI Architecture]

# Installpip install "langcare-cli @ git+https://github.com/langcare/langcare-mcp-fhir.git#subdirectory=cli"
# Uselangcare fhir search Patient --query "name=John"langcare fhir read Patient 123langcare fhir create Observation --data @obs.jsonlangcare fhir update Patient 123 --data @patient.json

The 40+ clinical skills in the Skills Library work as-is — skills reference abstract tool names, not transport. Register the CLI as subprocess tools in your agent framework and skills run without modification.

Full documentation and setup guide

Documentation

Getting Started

• 📖 Local Development & Testing Guide - Complete guide for local setup and testing
• 🚀 Installation & Configuration - Quick setup guide above

Agent Integration

• 🤖 Agent Prompt Guide - Complete guide for AI agents using LangCare MCP FHIR (tool examples, workflows, best practices)

Security & Authentication

• 🛡️ Security Documentation - Complete security architecture and HIPAA compliance
• 🔐 EPIC Setup Guide - JWT authentication, key generation, and JWKS registration
• 🔐 OpenEMR Setup Guide - SMART on FHIR Backend Services (private_key_jwt/RS384) setup, JWKS generation, and OpenEMR API client registration
• 📋 EPIC Scopes Reference - Complete OAuth2 scopes guide for FHIR resources
• 🔑 Authentication Methods - Supported auth methods

Deployment

• Fly.io Deployment Guide - Remote Streamable HTTP deployment, provider configs, secrets, Docker

Development & Testing

• 🧪 Testing Methods - Claude Desktop, MCP Inspector, manual testing, and automation
• 📦 Project Structure - Directory layout and architecture
• 🔧 Build Commands - Development workflow

Dependencies

• github.com/modelcontextprotocol/go-sdk - Official MCP SDK
• gopkg.in/yaml.v3 - Configuration parsing
• golang.org/x/oauth2 - OAuth2 client library
• github.com/golang-jwt/jwt/v5 - JWT signing and verification
• Go 1.25+

HIPAA Compliance

• PHI scrubbing enabled by default
• Never logs patient identifiers
• TLS support for HTTP transport
• Proper error sanitization
• Audit logging ready
• Stateless proxy design (no persistent storage)

Testing

Public Test Server

Default configuration uses HAPI FHIR public test server (https://hapi.fhir.org/baseR4) for immediate testing without setup.

Test Your Setup

• 📖 Local Development & Testing Guide - Complete guide for setup, testing with Claude Desktop, MCP Inspector, and automation
• 🔐 EPIC Security Setup - Detailed EPIC authentication guide
• 🛡️ Security Documentation - Production deployment and security

Claude Managed Agents

9 production-ready clinical AI agents built on the Anthropic Managed Agents API. Each agent connects to a LangCare MCP FHIR Server and uses a curated set of domain-specific clinical skills drawn from the 40+ Clinical Skills Library. Sessions are persistent, visible at platform.claude.com/workspaces/default/sessions, and can be run interactively or driven by a single prompt.

[LangCare — Claude Managed Agents]

Quickstart

# 1. Set environment variablesexport ANTHROPIC_API_KEY=sk-ant-...export LANGCARE_MCP_URL=https://langcare-mcp-dev.fly.dev/mcpexport LANGCARE_MCP_TOKEN=your-bearer-token
# 2. Upload skills, create environment + vault, deploy all 9 agentscd cma/scripts./setup.sh dev
# 3. Run a session./run-session.sh <agent-id> <env-id> <vault-id> "Show active medications for patient ID d886a934-5568-42b3-9324-0f0b05fc018c"

setup.sh is idempotent — safe to re-run. At the end it prints the Environment ID and Vault ID needed for sessions.

Full guide: cma/README.md — env vars, all scripts reference, troubleshooting.

Contributing

We welcome contributions from healthcare professionals, developers, and informaticists!

There are three main ways to contribute:

1. Core MCP Server (Go Development)

• Bug fixes and performance improvements
• New FHIR provider implementations (AllScripts, Athenahealth, etc.)
• Security enhancements and observability features
• Testing and CI/CD improvements

2. Clinical Skills (Healthcare Workflows)

• Evidence-based clinical workflows using FHIR
• Specialty-specific protocols (cardiology, oncology, etc.)
• Population health and quality measure workflows
• Clinical decision support algorithms

Skills are agent-agnostic workflow guides that work across Claude, ChatGPT, and Gemini. No coding required - just clinical expertise and FHIR knowledge!

3. MCP Apps (Interactive UIs)

• New clinical or administrative UI apps
• Enhancements to existing apps (FHIR Explorer, Patient Chart Review)
• Reusable components and patterns for healthcare UIs

See apps/README.md for the development guide.

4. Agent Integrations (Platform Setup)

• Setup guides for new AI platforms
• Deployment examples (Docker, Kubernetes, cloud)
• Monitoring and observability setups
• CI/CD pipelines

Get started: Read CONTRIBUTING.md for detailed guidelines, code standards, and submission process.

Recognition: Contributors are credited in README, release notes, and skill/integration author credits. Outstanding contributors may be invited as maintainers.

Questions? Open a GitHub Discussion or issue!

Community

• GitHub Discussions - Ask questions, share ideas: https://github.com/langcare/langcare-mcp-fhir/discussions
• GitHub Issues - Report bugs, request features: https://github.com/langcare/langcare-mcp-fhir/issues
• Contributing Guide - How to contribute: https://github.com/langcare/langcare-mcp-fhir/blob/main/CONTRIBUTING.md
• Skills - Clinical workflows: https://github.com/langcare/langcare-mcp-fhir/blob/main/skills/README.md

License

See LICENSE file.

Built with ❤️ by the LangCare team and contributors.

Improving healthcare through better AI infrastructure.