Inferensys

Glossary

Mirth Connect

Mirth Connect is a widely adopted open-source interface engine used for bi-directional sending, receiving, and transforming of HL7 v2 messages and other healthcare data formats across clinical systems.
Developer reviewing semantic search engine results on laptop, relevance scores visible, technical search demo.
INTERFACE ENGINE

What is Mirth Connect?

Mirth Connect is a widely adopted, open-source interface engine used for bi-directional sending, receiving, and transforming of HL7 v2 messages and other healthcare data formats across clinical systems.

Mirth Connect is an open-source integration engine specifically designed for healthcare interoperability. It acts as a central translation broker, enabling bi-directional message routing, filtering, and complex data mapping between disparate systems using standards like HL7 v2, FHIR, C-CDA, and DICOM. Its channel-based architecture allows developers to build granular data flows.

Deployed as a hub-and-spoke model, Mirth Connect drastically reduces the complexity of point-to-point interfaces by normalizing data into a canonical format. It provides guaranteed delivery mechanisms and a dead letter queue for failed transactions, ensuring no clinical message is lost during transmission between an EHR and a lab system.

INTERFACE ENGINE CAPABILITIES

Key Features of Mirth Connect

Mirth Connect (now NextGen Connect) is a cross-platform interface engine that provides a unified hub-and-spoke architecture for bi-directional healthcare data exchange. It specializes in HL7 v2.x message ingestion, transformation, and routing, while supporting a wide array of other healthcare standards and protocols.

01

Bi-Directional HL7 v2.x Transformation

Mirth Connect provides a robust pre-processor and post-processor scripting environment for parsing and manipulating pipe-and-hat delimited HL7 v2.x messages. It can map fields between disparate versions (e.g., converting an ADT^A04 from v2.3 to v2.5) using JavaScript or Python. The engine handles ACK/NAK generation natively, ensuring guaranteed delivery semantics by committing messages to a persistent queue until a valid application-level acknowledgment is received from the downstream system.

HL7 v2.1-2.8
Version Support
02

Multi-Protocol Connectivity

Beyond HL7 v2, Mirth Connect acts as a universal translator for clinical data formats. It supports FHIR R4 resource parsing, CCDA/CDA XML document processing, DICOM image routing, and X12 EDI claims transactions. For transport, it offers listeners and senders for MLLP, TCP/IP, HTTP(S), SFTP, and JMS. This allows a single engine instance to receive a lab result via ASTM, transform it into a FHIR Observation resource, and POST it to a cloud API.

15+
Supported Protocols
03

JavaScript-Based Filtering & Routing

Channel-level logic is written in JavaScript (Rhino engine), allowing for complex, programmatic message routing without external services. Developers can implement conditional logic to inspect message segments (e.g., msg['PV1']['PV1.3']['PV1.3.1'].toString()) and dynamically route to different destinations. This enables content-based routing where a single inbound channel can fan out messages to a lab system, a billing system, and a public health registry based on the MSH.9 message type or OBR.4 universal service identifier.

JavaScript
Scripting Engine
04

Graphical Channel Builder

Mirth Connect provides a drag-and-drop visual channel editor that abstracts the complexity of integration logic. Administrators configure Source (inbound listener), Destinations (outbound senders), and Filters/Transformers in a single pane. The interface displays real-time message statistics, including Received, Filtered, Queued, Sent, and Errored counts. This visual dashboard allows integration engineers to quickly debug message flow without tailing server logs, reducing mean time to resolution for interface failures.

Real-time
Message Dashboard
05

Guaranteed Delivery & Queuing

To ensure zero data loss in clinical contexts, Mirth Connect implements a persistent message queue with configurable retry logic. If a downstream EHR system is unreachable, messages are serialized to an embedded Apache Derby or external PostgreSQL database. The engine will retry delivery with exponential backoff until the message is successfully acknowledged. Unprocessable messages are routed to a Dead Letter Queue for manual intervention, ensuring that a network outage never results in a lost lab result or patient admission.

Zero Loss
Delivery Guarantee
MIRTH CONNECT BASICS

Frequently Asked Questions

Clear, technically precise answers to the most common questions about the architecture, capabilities, and operational mechanics of the NextGen Connect (formerly Mirth Connect) integration engine.

Mirth Connect is an open-source, cross-platform interface engine that facilitates the bi-directional sending, receiving, and transformation of healthcare data, primarily HL7 v2.x messages, between disparate clinical information systems. It operates on a hub-and-spoke model, acting as a central broker where channels—the core processing units—are configured. Each channel consists of a source connector that listens for incoming data on a specific protocol (such as LLP/MLLP for HL7, TCP, HTTP, or a file reader), a filter and transformer chain that manipulates the message payload using JavaScript or Python, and a destination connector that transmits the processed data to the target system. The engine parses incoming pipe-and-hat delimited HL7 messages into an internal XML representation, allowing developers to write procedural code to map fields, enrich data, and route messages based on trigger events like ADT^A01 or ORU^R01.

Prasad Kumkar

About the author

Prasad Kumkar

CEO & MD, Inference Systems

Prasad Kumkar is the CEO & MD of Inference Systems and writes about AI systems architecture, LLM infrastructure, model serving, evaluation, and production deployment. Over 5+ years, he has worked across computer vision models, L5 autonomous vehicle systems, and LLM research, with a focus on taking complex AI ideas into real-world engineering systems.

His work and writing cover AI systems, large language models, AI agents, multimodal systems, autonomous systems, inference optimization, RAG, evaluation, and production AI engineering.