ModMed (short for Modernizing Medicine) is a cloud-based electronic health record (EHR) and practice management company built for specialty medical practices. Its flagship EMA EHR is designed to adapt to specialty-specific exam workflows in fields like dermatology, ophthalmology, orthopedics, ENT, gastroenterology, urology, plastic surgery, and podiatry, and the broader ModMed suite also includes practice management and billing tools, the gGastro system for GI practices, the Klara patient engagement platform, ModMed Pay, and AI-powered documentation tools. More than 40,000 providers, ranging from solo practices to large multi-site groups, run on the platform.

The Model Context Protocol (MCP) is an open standard, developed by Anthropic, that lets AI models and AI agents connect to external tools, databases, and APIs in a standardized way. Instead of every AI product needing a custom integration for every software it wants to use, an MCP server exposes a system's data and actions through a common interface that any MCP-compatible AI assistant can call. In practice, this means an MCP server could let an AI assistant search patient records, check appointment schedules, or pull billing data from a system like ModMed, using the same protocol it would use with other tools.

Learn more: How do MCP servers work?

 

Does ModMed have an MCP server?

No, as of now, ModMed does not offer an official, first-party MCP server. ModMed has invested in developer connectivity, but that investment has gone into traditional API infrastructure rather than a native MCP implementation.

 

What ModMed offers

ModMed's own developer portal describes itself as the place to "begin discovering the ways which outside systems can connect to ModMed systems and customers," and states that ModMed "offers a Certified FHIR API including SMART on FHIR capabilities, a Proprietary API, more traditional HL7 options, and plenty of other options for connectivity."

The connectivity strategy centers on a few options:

  • EMA Certified FHIR API - FHIR (short for Fast Healthcare Interoperability Resources) is an industry-wide standard format for exchanging health data, meant to make different medical software systems able to understand each other. An API (Application Programming Interface) is the set of rules that lets one piece of software request data from another. So this is an HL7 FHIR-based interface, HL7 being the standards organization behind FHIR, with SMART on FHIR support, a related standard that lets outside apps securely plug into an EHR the way apps plug into a phone's app store. It's aimed at standards-based interoperability, letting patient-facing apps pull things like recent lab results.
  • EMA Proprietary API - "proprietary" here means it's ModMed's own custom-built version, rather than a generic industry standard. It's a customized FHIR R4 implementation (R4 is simply the current version number of the FHIR standard), described in ModMed's own docs, that goes beyond the certified API to support CREATE, READ, SEARCH, and UPDATE operations which are the basic actions of adding, viewing, searching for, and editing data across patients, appointments, and clinical data. It's secured with OAuth2 (a method for granting an app permission to access an account without handing over a password) and backed by a sandbox for testing, meaning a safe practice environment where developers can try things out without touching real patient data.
  • Traditional HL7 interfaces - HL7 (Health Level Seven) is an older, widely-used family of messaging standards that healthcare systems have used to send data like lab results or admission records back and forth.

 

Where the MCP confusion comes from

A search for "ModMed MCP" provides a few sources worth noting:

  1. Speculative blog content. Some articles, like "What Is ModMed MCP?" from GetGuru, explore what an MCP integration could look like for ModMed, walking through hypothetical scenarios for data accessibility and workflow automation.
  2. Third-party MCP platforms. Services like Keragon Healthcare MCP have built MCP servers that connect to a different EHR systems including ModMed, Athenahealth, Elation Health, Healthie, and others, as part of a HIPAA compliant automation layer, and Keragon also offers a dedicated ModMed integration page for its no-code workflow builder. This is a working MCP integration, but it's built and operated by a third party sitting on top of ModMed's existing APIs, not something ModMed ships or maintains itself.

What this means for practices

If a practice wants an AI assistant to interact with ModMed data, there are two paths:

  • Build your own connection. Use ModMed's FHIR or Proprietary API to pull data from ModMed, then wrap that connection in a small MCP server you build yourself. This lets AI tools like Claude use ModMed data.
  • Use a third-party MCP platform that has already built ModMed connectivity, accepting that you're adding another vendor and another layer of data access to manage.

Either way, it's worth keeping an eye on ModMed's developer portal and the synapSYS Marketplace, since specialty EHR vendors are moving quickly on AI connectivity, and an official MCP offering could arrive as the ecosystem matures.

Read also: Introducing the Paubox MCP server

 

FAQs

What is an EHR?

An EHR (electronic health record) is digital software that stores a patient's medical history, treatment records, and clinical notes in place of paper charts.

 

Is MCP only useful for healthcare software?

No, MCP is a general-purpose protocol used to connect AI assistants to all kinds of software, from project management tools to e-commerce platforms.

 

Is it safe to connect AI tools to patient data?

Any AI integration touching patient data needs to be handled carefully, with proper security measures like encryption and HIPAA compliant data handling agreements in place.

 

What's the difference between an API and an MCP server?

An API is the set of rules a system uses to share data, while an MCP server is a standardized wrapper built on top of one or more APIs designed for AI assistants to use.