If you work anywhere near healthcare data in the United States, you have almost certainly either used Mirth Connect or sent a message through an interface that did. Since 2006 it has quietly become the default choice for moving clinical data between systems — used by hospitals, reference labs, radiology centers, payers, HIEs, and nearly every mid-market healthtech startup that needs to exchange data with an EHR.
This pillar guide walks through what Mirth Connect is, how it works, what it costs, where it shines, where it struggles, and when you should bring in outside expertise. It links out to deeper technical guides in our knowledge base for engineers who want to go further.
1. What Is Mirth Connect?
Mirth Connect is an open-source, cross-platform healthcare integration engine built on Java. Its job, in one sentence, is to accept a healthcare message in one format, transform it, and deliver it to one or more destinations in another format — reliably, securely, and at scale.
In practice, that job looks like this:
- A hospital's admission system sends an HL7 v2 ADT^A01 message over MLLP when a patient is admitted.
- Mirth Connect receives it, parses the segments, optionally transforms the data (for example mapping the hospital's facility code to the lab's expected code), and writes it to a database, a REST endpoint, a file share, or another HL7 listener.
- An acknowledgment (
ACK) is returned to the sender within milliseconds. - Every message, every transformation, and every error is logged and available for audit.
Mirth Connect does that hundreds of times per second across tens of thousands of hospitals worldwide. For a plain-English introduction aimed at non-technical stakeholders, see our separate walk-through: What Is Mirth Connect — a Guide for Healthcare Teams.
Key facts at a glance
| First released | 2006 |
| Current product name | NextGen Connect Integration Engine |
| License | Mozilla Public License 2.0 (open source) |
| Language | Java (server), with JavaScript + Groovy for scripting |
| Database backends | PostgreSQL, MySQL, Oracle, Microsoft SQL Server, Apache Derby |
| OS support | Windows, Linux, macOS; runs in Docker/Kubernetes |
| Deployment size | From single-server labs to multi-region HIE clusters |
| Commercial support | NextGen Healthcare (official) and third-party partners like Taction Software |
2. A Short History: From Mirth Corporation to NextGen Connect
Understanding the product's history explains some of the terminology confusion you'll encounter in documentation, forums, and job descriptions.
- 2006 — Mirth Corporation releases the first version of Mirth Connect as open-source software.
- 2013 — Quality Systems, Inc. (QSI), parent company of NextGen Healthcare, acquires Mirth Corporation.
- 2019 — QSI rebrands NextGen Healthcare and officially renames Mirth Connect to NextGen Connect Integration Engine.
- 2020–present — NextGen continues to release the open-source core (still branded "Mirth Connect" in the installer and codebase) alongside a paid commercial tier.
In practice, most of the industry still calls the product Mirth Connect or just Mirth. Job postings, books, Stack Overflow threads, and internal documentation almost never say "NextGen Connect." When you see either name, they refer to the same product.
3. How Mirth Connect Works: Core Architecture
At a high level, Mirth Connect is a message broker purpose-built for healthcare data. Its architecture has three main layers.
3.1 The Server
The Mirth Connect server is a Java application that runs as a background service. It hosts:
- A message engine that receives, queues, transforms, and dispatches messages.
- An embedded web server (Jetty) that serves the Administrator client and optional REST APIs.
- A database layer that stores channel configurations, messages, logs, and metadata.
- Java and Rhino JavaScript runtimes for running custom scripts inside message processing.
3.2 The Administrator (Client)
A Java Swing desktop application (and, in recent versions, a browser-based option) that engineers use to:
- Build and configure channels.
- Deploy, start, stop, and redeploy channels.
- View messages, logs, alerts, and metrics.
- Manage users, roles, permissions, and global scripts.
3.3 The Database
Mirth stores everything — configuration, every message, every attempt, every error — in a relational database. That makes incident investigation and compliance reporting straightforward, but it also means database sizing and tuning is the single most common production issue in Mirth deployments.
4. Channels, Connectors, Filters, and Transformers
Everything in Mirth Connect happens inside a channel. Understanding the four moving parts inside each channel is the single most important concept to grasp.
4.1 Channels
A channel is a pipeline that defines:
- Where a message comes from (the source connector)
- What transformations are applied
- Where it goes (one or more destination connectors)
- Business rules for routing, filtering, retrying, and handling errors
A production hospital deployment typically runs anywhere from 20 to 500 channels simultaneously.
4.2 Connectors
The most common connector types:
| Connector Type | Typical Use |
|---|---|
| TCP / MLLP Listener | Receiving HL7 v2 messages from EHRs and clinical systems |
| TCP / MLLP Sender | Sending HL7 v2 to downstream systems |
| HTTP Listener / Sender | REST APIs, FHIR endpoints, webhooks |
| Database Reader / Writer | Reading from or writing to SQL databases |
| File Reader / Writer | Processing CSV, flat files, CDA XML |
| SFTP / FTPS | Scheduled file exchanges with payers and labs |
| JMS Listener / Sender | Enterprise message buses |
| Web Service (SOAP) | Legacy integrations |
| JavaScript Reader / Writer | Custom source/destination logic |
4.3 Filters
A filter is a rule that decides whether a message should be processed or dropped. For example: "Only process ADT messages where MSH-4 (sending facility) equals MEMORIAL_HOSPITAL." Filters can be written in JavaScript, Groovy, or using rule builders.
4.4 Transformers
A transformer modifies the message. This is where most of the real work happens. Common transformations:
- Mapping HL7 v2 fields to FHIR resources (and vice versa).
- Converting date formats, code sets, and identifiers.
- Adding segments (e.g., inserting a
ZPIsegment for custom patient identifiers). - Stripping PHI for downstream analytics.
- Converting JSON to XML, or flat files to HL7.
Transformers use either JavaScript (Mozilla Rhino) or Groovy. The choice matters for performance and developer ergonomics.
5. Supported Healthcare Data Standards
Mirth Connect's killer feature is that it speaks every major healthcare data standard out of the box.
5.1 HL7 v2
Versions 2.1 through 2.8 are supported natively. Every HL7 v2 message type you're likely to encounter — ADT, ORM, ORU, SIU, MDM, DFT, RDE, VXU, ACK — can be parsed, validated, and generated. For deeper coverage, see our HL7 integration services.
5.2 HL7 v3 and CDA
HL7 v3 XML messages and Clinical Document Architecture (CDA) R2 documents — including C-CDA templates used for Meaningful Use / Promoting Interoperability — are handled via XML parsers and XSLT transformers.
5.3 FHIR
FHIR DSTU2, STU3, and R4 are supported through HTTP connectors and JSON/XML processing. Mirth Connect is a popular choice for building FHIR façades in front of legacy HL7 v2 systems — a pattern we walk through in Mirth Connect + FHIR: Powering the Next Generation of Healthcare Data Exchange. If your team needs a production FHIR server to pair with Mirth, see our Mirth FHIR Server offering.
5.4 X12 EDI
Mirth can process X12 transactions (270/271 eligibility, 834 enrollment, 835 remittance, 837 claims) via dedicated parsers or configured delimiters.
5.5 DICOM
DICOM C-STORE, C-FIND, and WADO integrations are possible using custom connectors and the open-source dcm4che toolkit. Typical use cases are routing imaging orders and results.
5.6 NCPDP
NCPDP SCRIPT and Telecommunication Standard messages for e-prescribing can be handled using the flat-file and custom transformer approach.
5.7 Custom and proprietary formats
Because connectors are extensible and transformers run arbitrary code, Mirth Connect can adapt to virtually any format — including vendor-proprietary formats, legacy flat files, and binary protocols.
6. Common Use Cases
Knowing what Mirth Connect does is less useful than knowing what people actually use it for. Here are the deployment patterns we see most often.
6.1 Hospital EHR integration hub
The most common pattern. Mirth sits between the hospital EHR (Epic, Cerner, Meditech, etc.) and dozens of ancillary systems. See our dedicated guides on Epic EHR integration and Cerner EHR integration for the specifics.
6.2 Laboratory interface engine
Reference labs and hospital labs use Mirth to receive orders (ORM^O01), route them to LIMS, and send results (ORU^R01) back to hundreds of ordering providers. A typical regional lab can push 1–5 million HL7 messages per day through a single Mirth cluster.
6.3 Healthtech startup FHIR façade
Modern digital health companies typically need to pull data from a customer's EHR in real time. Mirth Connect converts the hospital's legacy HL7 v2 feed into FHIR R4 resources the startup's app can consume — without requiring the hospital to modernize.
6.4 HIE and population health
Health information exchanges use Mirth to normalize data from dozens of participating organizations, deduplicate patient records, and route queries and subscriptions.
6.5 Payer-provider data exchange
Mirth processes X12 EDI transactions between payers and providers — eligibility checks, claims, and remittances — often alongside HL7 and FHIR.
6.6 Compliance and reporting
Mandatory reporting — immunization registries (IIS), cancer registries, syndromic surveillance to CDC, prescription drug monitoring (PDMP) — is typically handled by one or two dedicated Mirth channels per program.
7. Deployment Options: On-Premises, Cloud, Containerized
Mirth Connect runs anywhere the JVM runs. The three dominant deployment models are:
7.1 On-premises (bare metal or VM)
Still the most common in hospital environments, where PHI never leaves the datacenter. Typical setup: one or two application servers in an active/passive HA pair, a dedicated database server, and a load balancer in front for MLLP and HTTPS.
7.2 Cloud (AWS, Azure, GCP)
Increasingly the default for healthtech startups and modernizing health systems. Mirth runs on EC2, Azure VMs, or GCE, with managed databases (RDS, Azure SQL, Cloud SQL).
7.3 Containerized (Docker / Kubernetes)
Newer deployments often run Mirth in Docker containers orchestrated by Kubernetes, with Helm charts for templated deployment. This gives you rolling upgrades, declarative configuration, and horizontal scaling.
Choosing between them
| Factor | On-Prem | Cloud | Kubernetes |
|---|---|---|---|
| PHI stays in your datacenter | Yes | BAA required | BAA required |
| Upfront cost | High | Low | Medium |
| Scaling flexibility | Low | High | High |
| Operational complexity | Medium | Medium | High |
| DR / HA out of the box | No | Yes (partial) | Yes |
| Ideal for | Established hospitals | Most healthtech | High-scale platforms |
8. Installation Overview
A basic Mirth Connect installation takes 30–60 minutes on a clean server. The broad steps:
- Install Java (OpenJDK 11+ for current versions).
- Install a database (PostgreSQL is recommended for production).
- Download the Mirth Connect installer from the NextGen GitHub releases page.
- Run the installer and set the web admin port, database connection, and keystore password.
- Install the Administrator client on your workstation.
- Configure SSL/TLS for the admin, API, and listener ports.
- Create your first channel — typically an HL7 file-to-file test channel.
That's the 30-second version. The full production-grade walkthrough with screenshots, security hardening, and database tuning is in our Mirth Connect Installation Guide for Windows & Linux.
9. Licensing: Open Source vs Commercial
This is one of the most common questions we get from buyers.
Open-source edition (Mirth Connect)
- Cost: Free
- License: MPL 2.0
- Support: Community forums, Stack Overflow, third-party partners
- Updates: Published to GitHub
- Limitations: No official SLA, no certified binaries, no commercial LDAP/SSO modules in some versions
Commercial edition (NextGen Connect Integration Engine)
- Cost: Licensed annually by NextGen Healthcare — typically $20K–$80K/year depending on scale and modules
- Support: NextGen's official support team
- Updates: Certified binaries with patched CVEs
- Extras: Advanced commercial connectors, priority bug fixes, official documentation
Which should you pick?
Most of our clients run the open-source edition with third-party expert support. They get feature parity with the open product, vendor-neutral engineering help with faster SLAs than NextGen typically offers, and a far lower total cost of ownership. Our Mirth Connect support services are designed for exactly this model — production-grade 24/7 coverage without the commercial license cost.
10. Mirth Connect vs Alternatives
Mirth Connect is the market leader, but it's not the only option. The serious alternatives in US healthcare are:
| Product | Maker | Typical Buyer | Starting License |
|---|---|---|---|
| Rhapsody | Lyniate | Large IDNs, HIEs | $$$ (~$50K+/yr) |
| Iguana | iNTERFACEWARE | Mid-market hospitals, labs | $$ |
| InterSystems HealthShare / IRIS | InterSystems | Large enterprises, HIEs | $$$$ |
| Corepoint Integration Engine | Lyniate (ex-NeoTool) | Hospitals | $$$ |
| Cloverleaf | Infor | Legacy hospital customers | $$$ |
| Mirth Connect | NextGen Healthcare | Everyone | Free (OSS) |
The short version: Mirth wins on cost and flexibility; the others typically win on hand-holding, vendor support, and polish.
11. Common Challenges (and How to Solve Them)
The Mirth Connect learning curve is real, and production deployments hit predictable pain points. The top five:
11.1 Java heap exhaustion and OutOfMemoryError
Under-sized JVM heap is the most common cause of channel failures. High-throughput channels that buffer large messages (CDA documents, base64-encoded images in HL7) need careful -Xms/-Xmx tuning.
11.2 Database bloat
Every message stored in Mirth's database by default. Without a retention and pruning policy, the database can grow hundreds of gigabytes per month. Configure channel pruning from day one.
11.3 MLLP connection failures
MLLP over TCP is famously finicky — firewalls, proxies, and TLS mismatches all cause the same-looking error.
11.4 Transformer performance
JavaScript transformers with synchronous regex or external HTTP calls can slow a channel from 1,000 msg/sec to 10 msg/sec. Profiling transformers is a critical skill.
11.5 Version upgrades
Upgrading from Mirth 3.x to 4.x (NextGen Connect) involves API changes in transformers, database schema migrations, and regression risk. Most teams under-budget the testing effort. We've compiled a full list of common Mirth Connect issues and fixes from production environments.
12. Security and HIPAA Compliance
Mirth Connect handles PHI, which means HIPAA applies— and if you are a business associate, so do the HIPAA Security Rule's administrative, physical, and technical safeguards.
The non-negotiables for any production Mirth deployment:
- TLS 1.2+ on every listener and sender, with strong cipher suites and current certificates.
- Separate keystores for admin, API, and channel traffic.
- Named user accounts with role-based access control — no shared
adminlogins. - Audit logging enabled and shipped to a SIEM (Splunk, Elastic, Sumo Logic).
- Database encryption at rest (transparent data encryption or volume encryption).
- Channel-level PHI minimization — drop fields you don't need downstream.
- Regular key rotation and a documented credential inventory.
- Patching cadence — quarterly at minimum, same-day for critical CVEs.
- Penetration testing of the admin API and any HTTP listeners.
13. When to Bring in Expert Support
Mirth Connect is free, but running it safely at scale is not. The point at which most teams call for help:
- A production channel outage is affecting clinical workflows and no one on the team has seen the error before.
- A critical interface migration (Mirth 3.x → 4.x, or cloud migration) is on a hard deadline and the team has never done one.
- An EHR go-live (Epic, Cerner, athenahealth) where HL7 or FHIR interface build is on the critical path.
- HITRUST, SOC 2, or HIPAA audit remediation flagged integration-layer gaps.
- FHIR R4 certification (ONC Health IT Certification Program) is needed for a product launch.
- The team lost its senior HL7 engineer and needs coverage while recruiting.
Our Mirth Connect helpdesk is built exactly for these scenarios — 24/7 emergency response with an under-15-minute SLA for critical channel failures, staffed by engineers with 12+ years of HL7 and FHIR experience. For longer-term engagements, our full services portfolio covers everything from greenfield HL7 interface development to full HL7 integration programs across the USA.
14. Frequently Asked Questions
Is Mirth Connect free?
Yes. The open-source Mirth Connect Integration Engine is released under the Mozilla Public License 2.0 and is free to download, use, and modify. NextGen Healthcare also sells a commercial edition (NextGen Connect Integration Engine) with official support and certified binaries.
Is Mirth Connect the same as NextGen Connect?
Yes. NextGen Connect Integration Engine is the rebranded name for Mirth Connect after NextGen Healthcare's acquisition in 2013 and official rebrand in 2019. Most of the industry still uses the name "Mirth Connect."
What is the latest version of Mirth Connect?
NextGen Connect Integration Engine is updated regularly. The 4.x branch is the current supported line, with the open-source project publishing releases on GitHub. Always confirm the latest stable version on the official releases page before installing.
Can Mirth Connect handle FHIR?
Yes. Mirth Connect supports FHIR DSTU2, STU3, and R4 through HTTP connectors and JSON/XML processing. It is widely used as a FHIR façade in front of legacy HL7 v2 systems.
What databases does Mirth Connect support?
PostgreSQL, MySQL, Oracle, Microsoft SQL Server, and Apache Derby (for development only). PostgreSQL is the most common choice for production deployments.
How many messages per second can Mirth Connect handle?
A properly tuned single-node Mirth Connect server commonly handles 500–2,000 messages per second for typical HL7 v2 workloads. Clustered deployments scale linearly — the largest installations process tens of millions of messages per day.
Is Mirth Connect HIPAA compliant?
Mirth Connect is software, not a service — software itself cannot be "HIPAA compliant." However, Mirth Connect provides all the technical controls (TLS, access control, audit logging, PHI-aware transformers) needed to build a HIPAA-compliant integration environment when deployed and operated correctly.
Do I need a Mirth Connect consultant?
Not for simple development or a single-interface pilot. You typically need expert help when you have critical production workloads, tight timelines, complex multi-system migrations, compliance pressure, or no senior HL7 engineer on staff.