Mirth, Rhapsody or Iguana: the short answer
There is no overall winner. Mirth Connect, now licensed per server by NextGen, and its free fork the Open Integration Engine suit teams with JavaScript skills who want to own and run the engine. Rhapsody suits large health systems and HIEs that want one commercial vendor, wide standards coverage and a hosted option. Iguana suits teams who like scripting in Lua with a small, direct vendor. Choose on the skills you can hire, hosting, support and five-year cost, not licence price alone.
A comparison from people who mostly work on one of the engines. Read it with that in mind
Most of our interface work is on Mirth Connect and the Open Integration Engine, so we know those two far better than the others, and we have a commercial interest in projects that use them. To keep this page useful anyway, every factual row below comes from a vendor’s own page, release notes or a government advisory, dated, and linked. Where a vendor publishes nothing, such as a price list, the table says so instead of guessing. We are not a partner or reseller of NextGen, Rhapsody, iNTERFACEWARE, InterSystems, AWS, Microsoft or Google.
The choice matters more than it used to. NextGen’s move to a commercial-only Mirth in 2025 sent many open-source users looking, Rhapsody and iNTERFACEWARE have both changed their product line-ups, and the cloud platforms now handle parts of the job that once needed an engine. The answer for a 40-interface lab network is rarely the answer for a statewide HIE.
If you already know you are staying in the Mirth family, go straight to our Mirth Connect migration guide. If you are still choosing, the table, the checklist and the cost section below are the parts to share with your team. We can also run the selection for you as a fixed-scope piece of work, described further down, and support the result through our interface engine and Mirth Connect services.
How we compared them
We read each vendor’s current public material in the first week of November 2026: product pages, brochures, release notes, licence pages and, for Mirth, the CISA advisory. We did not run benchmarks, and we do not quote throughput figures, because vendor numbers are not comparable with each other. Rhapsody’s documentation and NextGen’s user guide sit behind logins, so for those two some detail comes from brochures and public release notes. Check anything decisive with the vendor in writing.
The four engines side by side
| Mirth Connect (NextGen) | Open Integration Engine | Rhapsody | Iguana (iNTERFACEWARE) | |
|---|---|---|---|---|
| Who makes it | NextGen Healthcare | Community maintainers and an elected steering committee; Eclipse Foundation incubating project since September 2026 | Rhapsody, Boston (renamed from Lyniate in 2023; Hg bought the Rhapsody business from Orion Health), which also sells the Corepoint engine | iNTERFACEWARE, a small company building integration software since 1997 |
| Current release | 4.7 line; CISA names 4.7.2 or later as fixed for the 2026 flaws | 4.7.0, 21 October 2026 | Not published openly; documentation needs a login | Iguana 6.2 and IguanaX 10.1.119 offered side by side; 6.3 in progress |
| Licensing | Commercial since 4.6; annual fee per server, not per interface; Enterprise, Gold and Platinum tiers; free licence for non-profit research studies | Open source (MPL 2.0), no fee; paid support from independent vendors if you want it | Commercial | Commercial; terms agreed with each organisation |
| Published price | None | Free | None found | None; terms agreed per customer |
| How you build interfaces | Channels with filters and transformers in JavaScript (Rhino), able to call Java; reusable code templates | Same model as Mirth; imports Mirth channel exports | Graphical routes and filters; JavaScript filters and a mapper that generates JavaScript; a development kit for custom components | Scripts in Lua in the Translator editor, with message definition files for HL7 |
| Standards | HL7 v2 and v3, X12, NCPDP, DICOM, XML, JSON, delimited; FHIR through a commercial extension | As Mirth 4.5.2, without NextGen’s commercial FHIR extension; FHIR handled with HTTP and JSON in code | HL7 v2 and v3, FHIR, CCDA, NCPDP, X12, IHE profiles, DICOM, XML, JSON (vendor brochure) | HL7 v2, X12, JSON, XML and web APIs, scripted in Lua |
| Where it runs | You host it: Windows, Linux, macOS, containers | You host it; container build added in 4.7.0 | Windows, Linux or AIX on your servers, or Rhapsody as a Service; Corepoint also as SaaS or iPaaS | Your servers: Windows, Linux, macOS |
| Monitoring | Dashboard and alerts; Mirth Command Center for environment metrics in all paid tiers | Dashboard and alerts; community plugins such as Sentinel for queue depth, error rate and inactivity alerts | Rhapsody Dashboard across environments, alerts, REST API for engine health | Web dashboard and logs per channel |
| High availability | Clustering through a commercial extension | No clustering in the core; a standby server and runbook | Vendor states it can be configured for availability | Ask the vendor; not described on its public pages |
| Admin interface | Desktop Administrator via launcher | Desktop Administrator; browser-based administrator plugin since August 2026 | Rhapsody IDE (a Windows application) and a web Management Console | Browser |
| Skills market (our view) | Widest: years as free software mean many integration engineers have used it | Same skills as Mirth | Specialists common in large health systems and HIEs, fewer and dearer elsewhere | Smaller pool; Lua is quick for developers to pick up |
| Typical buyer (our view) | Hospitals, labs and health-tech vendors wanting vendor backing for a familiar engine | Teams with in-house skills who want no licence fee and open code | Large health systems, HIEs and public bodies wanting one accountable vendor | Vendors and mid-sized providers who prefer code to configuration screens |
Sources: NextGen licensing announcement, NextGen Mirth product page, CISA ICSMA-26-253-01, Open Integration Engine, Rhapsody company facts, Corepoint, iNTERFACEWARE downloads and licensing approach. BridgeLink, Innovar Healthcare’s Mirth fork, sits close to the OIE column with commercial support from Innovar; it released 26.9.0 on 1 October 2026.
Also on the shortlist: InterSystems and the cloud platforms
| Option | What it is | What it handles | What you still build |
|---|---|---|---|
| InterSystems HealthShare Health Connect | Commercial engine on the IRIS data platform; 2026.1 (March 2026) added a Mirth migration tool, 2026.2 followed in July | HL7 v2, FHIR, X12 and DICOM, with graphical transformation and process editors and ObjectScript or Python for code | Interfaces, and the specialised skills to run it; strongest where InterSystems products already run |
| Google Cloud Healthcare API | Managed HL7v2, FHIR and DICOM stores | Receives HL7 v2 through an open-source MLLP adapter on GKE, parses and stores it, and publishes notifications | All routing and transformation logic, usually in your own pipeline |
| Azure Health Data Services | Managed FHIR service; the older Azure API for FHIR was retired on 30 September 2026 | $convert-data turns HL7 v2, C-CDA and JSON into FHIR R4 with Liquid templates | An MLLP listener, orchestration, error handling and replay |
| AWS | HealthLake (managed FHIR R4) and B2B Data Interchange (X12) | FHIR storage and search; X12 5010 HIPAA transactions; a CCDA-to-FHIR conversion agent announced in March 2026 | HL7 v2 reception and transformation, often with an engine running on AWS |
Sources: InterSystems 2026.1 announcement, Google HL7v2 concepts, Microsoft $convert-data and migration FAQ, AWS HealthLake features, AWS B2B Data Interchange. The pattern across all three clouds: they store and convert well, but none replaces an engine’s routing, queuing, retries and per-partner quirks on its own.
When each one fits
Mirth Connect under a NextGen licence
Pick it when you already run Mirth, your team writes JavaScript, and you want vendor-backed patches, Command Center and the commercial extensions without rewriting anything. Think twice when you run many small servers, because the fee is per server, or when the only reason is that nobody has time to evaluate the fork.
Open Integration Engine
Pick it when you have people who can run the engine themselves, you want open code and no licence fee, and your channels do not lean on NextGen’s paid extensions. Its 4.7.0 release is described by the project as recommended for production. Think twice when you need a single vendor to answer the phone at 3 a.m. and have no support contract with anyone; community support is generous but not a service level.
Rhapsody
Pick it when you are a large organisation with many partners and standards, you want one accountable vendor, a hosted or managed option and product features such as dashboards and EMPI or terminology add-ons from the same company. Think twice when your budget is small, your team is small, or you cannot easily hire people who know it. Corepoint, from the same vendor, is pitched at teams without a large IT department and is worth a separate look.
Iguana
Pick it when your developers prefer writing code to configuring screens, you value a small vendor you can talk to directly, and on-premises hosting is a requirement. Think twice when you need clarity on which generation to standardise on: iNTERFACEWARE currently supports both Iguana 6 and IguanaX and describes its direction as gradual convergence, so ask which one it recommends for a new installation and why.
InterSystems or a cloud platform
Pick InterSystems when you already run its products and have the skills. Pick a cloud service when your target is a FHIR store or analytics platform and an engine, or a small amount of code, can feed it. Pairing them is common: an engine handles HL7 v2 partners while the cloud service holds FHIR data. Our HL7 to FHIR conversion guide covers that middle step.
A decision checklist
- Interfaces: list every feed with message type, trigger events, volume, peak and partner. Count the ones that are not HL7 v2.
- Standards: do you need FHIR, X12, DICOM or IHE profiles inside the engine, or can another component handle them?
- Availability: how long can a feed be down before patients are affected? That sets your recovery time, and so whether you need clustering or just a tested standby.
- Hosting: on-premises only, your own cloud account, or the vendor’s hosting? Where must data stay?
- Skills: JavaScript, Lua, ObjectScript or graphical configuration. Can you hire for it where you are, and who covers holidays?
- Support: vendor contract, independent support firm, in-house only, or a mix? What response time do you actually need overnight?
- Monitoring: can it alert on queue depth, silent feeds and error rates, and send those alerts where your on-call team looks?
- Security and audit: how quickly does the vendor or project publish fixes? Mirth’s September 2026 advisory is a fair test question for any vendor: how would you have handled it?
- Cost over five years: see the next section. Ask every vendor for a written quote with renewal terms.
- Exit: can you export interface definitions in a readable form, and how hard would leaving be in five years?
Total cost of ownership: what to count
None of the commercial engines publishes a price list today, and we will not guess one. These are the lines to put in your model, with the vendor’s written quote in the first.
| Cost line | What drives it | How engines differ |
|---|---|---|
| Licence or subscription | Servers, environments, tiers, interfaces or volume, renewal uplift | Mirth charges per server, so test and disaster-recovery servers count; the fork is free; Rhapsody and Iguana quote per organisation |
| Hosting | Servers, database, storage for message history, backups, network links | Similar for self-hosted engines; a vendor-hosted option moves this into the subscription |
| Building interfaces | Number and complexity of feeds, partner testing | Driven by your team’s fluency more than by the engine |
| Running and support | On-call cover, incident handling, partner changes, upgrades | Vendor support is in the fee for commercial engines; for the fork you staff it or buy it |
| Upgrades and security fixes | Release cadence, testing effort, Java or OS changes | A regression test harness cuts this on any engine |
| Training and hiring | Local skills market, staff turnover | Cheapest where the skills are common in your market |
| Monitoring | Built-in features versus add-on tools | Some vendors bundle dashboards in paid tiers; open-source users add plugins or their own tooling |
| Exit | Interfaces to rebuild, partners to re-test | Lowest between Mirth and its forks, highest across vendor families |
What switching engines really costs
Between Mirth and the Open Integration Engine or BridgeLink, channels import directly, so the work is mostly testing, Java and extension replacement; our migration guide covers it. Between vendor families nothing imports cleanly: a Mirth JavaScript transformer becomes a Lua script or a Rhapsody route, rewritten and re-tested. Vendors offer help: Rhapsody markets an Intelligent Mapper that analyses legacy engines, and InterSystems added a Mirth migration tool in 2026.1. Treat these as accelerators, not as a replacement for testing with real messages.
Typical effort to rebuild and test one interface on a new engine, from our own estimates, confirmed per project:
- Simple feed (for example, ADT with light field mapping): 1–2 engineer-days.
- Typical clinical feed (lab orders and results with code tables): 3–6 engineer-days.
- Complex feed (several destinations, database look-ups, custom acknowledgements, payer-specific X12): 5–10 engineer-days.
Add partner coordination, a parallel run and staff retraining. For 60 interfaces that is a project of months, which is why the cheapest engine on paper is often the one you already run.
Illustrative scorecard for two invented organisations
Illustrative, not client work. Weights and scores are our judgement for invented profiles; use your own. Scores are 1 to 5, weighted to a total out of 5.
| Criterion | Weight A / B | Licensed Mirth | OIE | Rhapsody | Iguana |
|---|---|---|---|---|---|
| Fit with current team skills | 25 / 10 | 5 | 5 | 2 | 3 |
| Standards needed | 15 / 25 | 4 | 3 | 5 | 4 |
| Support model | 15 / 25 | 4 | 2 | 5 | 4 |
| Licence and running cost | 25 / 10 | 3 | 5 | 2 | 3 |
| Monitoring and availability | 10 / 20 | 4 | 3 | 5 | 3 |
| Cost to leave later | 10 / 10 | 4 | 5 | 2 | 3 |
| Total, profile A | 4.00 | 4.05 | 3.20 | 3.30 | |
| Total, profile B | 4.00 | 3.35 | 4.10 | 3.50 |
Profile A is a three-hospital group with two JavaScript-literate interface engineers, 60 mostly HL7 v2 feeds and a tight budget. Profile B is a regional HIE with many partners and standards, a 24-hour support requirement and appetite for a hosted service. The same engines, weighted differently, give opposite answers, and the margins are small. That is the point of doing the exercise rather than reading a ranking.
Where each product line stands this month
Product changes in the last year that should feature in any shortlist.
Mirth Connect 4.7 line
Commercial since 4.6, per-server annual fee. CISA advisory of 10 September 2026 names 4.7.2 or later as the fixed release.
Open Integration Engine 4.7.0
Released 21 October 2026; an Eclipse Foundation incubating project since September; browser administrator plugin since August.
Rhapsody, Corepoint and Axon
Two engines from one vendor; the Axon AI integration agent became generally available in March 2026 inside both.
Iguana 6 and IguanaX
Both generations supported; current downloads are Iguana 6.2 and IguanaX 10.1.119, with Iguana 6.3 adding Git push in progress.
Health Connect 2026.1 and 2026.2
The March 2026 extended-maintenance release added a Mirth migration tool; a continuous-delivery release followed in July.
Azure API for FHIR retired
Retired 30 September 2026; the FHIR service in Azure Health Data Services replaces it and offers $convert-data.
Facts, versions and vendor details checked 7 November 2026 against vendor sites, release notes and public advisories.
Help choosing an engine, and moving to it
Effort is given in typical engineer-days and confirmed as a fixed price per phase after a short scoping call. You deal with vendors directly for trials and quotes; we help you ask the right questions.
Engine selection assessment
The problem
You need to choose or confirm an engine, and the decision has to survive scrutiny from finance, security and clinical teams.
How we do it
We inventory your interfaces, agree weighted criteria with your team, prepare the questions for each vendor, build a five-year cost model around the quotes you receive, and write up a recommendation with the risks of each option. Typical effort: 5–10 engineer-days.
What you get
- Interface inventory and requirements matrix
- Vendor question set
- Weighted scorecard with your weights
- Five-year cost model and recommendation
Proof of concept on two engines
The problem
The scorecard is close and you want evidence, not brochures.
How we do it
The same two or three representative interfaces built on two shortlisted engines, using trial licences you obtain from the vendors, tested with de-identified samples, with notes on build effort, monitoring and operability. Typical effort: 8–15 engineer-days.
What you get
- Representative interfaces on each engine
- Test results with your samples
- Operability notes
- Revised recommendation
Migration to the chosen engine
The problem
The decision is made and now dozens of feeds have to move without interrupting care.
How we do it
Inventory, rebuild or import, replay-and-diff testing with production samples, a parallel run and cutover in waves. Within the Mirth family this is our core work; on other engines we work alongside your vendor or in-house team. Typical effort per interface is shown in the guide above.
What you get
- Migration plan and wave schedule
- Rebuilt or imported interfaces
- Comparison reports
- Cutover and rollback runbook
Running the engine afterwards
The problem
You want the new engine monitored and supported without hiring a full team.
How we do it
Monitoring for queue depth, silent feeds and errors, a support retainer sized to your interfaces and response times, and documented changes. Deepest for Mirth and the Open Integration Engine, as declared above.
What you get
- Monitoring and alert routing
- Support with agreed response times
- Change records and runbooks
- Quarterly review
How an engine selection assessment runs
About three to four weeks elapsed, mostly waiting on vendor answers.
Inventory and criteria
Interface list from your current engine or spreadsheets, and weighted criteria agreed with IT, security, finance and a clinical lead.
Vendor questions
You send our question set to the shortlisted vendors and request written quotes with renewal terms; we review answers as they arrive.
Scoring and cost model
Scorecard with your weights and a five-year cost model built from the quotes, with the assumptions written down.
Recommendation
A readout with the recommendation, the runner-up and what would change the answer, plus a migration outline.
Where our own products touch integration engines
Two products we build and run. Both explain the bias declared above: our tooling is built around the Mirth family.
Radiatus OneHealth
Certified OpenEMR plus 25 practice modules, with nothing to migrate Radiatus OneHealth includes an HL7 engine monitoring module for the Open Integration Engine, BridgeLink and Mirth Connect up to 4.5.2. An agent beside the engine reports channel states and message counts and raises alerts on queue depth, feed silence, error rate and JVM heap every two minutes. It does not monitor Rhapsody or Iguana. Visit openemr.radiatus.com
PathLab Software
Lab information system — sample to signed WhatsApp report Our laboratory information system takes analyser results over HL7 v2, ASTM or serial links through a connector on the lab PC, so result matching, flags and worklists are problems we handle in our own product, not only in client projects. Visit pathlabsoftware.comGuides and related pages
Guides
Glossary
Questions about Mirth, Rhapsody and Iguana
Neither is better in general. Mirth, or its free fork the Open Integration Engine, usually costs less and suits teams who write JavaScript and run their own servers. Rhapsody usually suits large organisations that want one accountable vendor, broad standards support and a hosted option. Weigh your skills, support needs and five-year cost.
It depends on your developers. Iguana scripts are written in Lua, which experienced programmers pick up quickly, while Mirth uses JavaScript, which more people already know. In our experience Mirth has the larger pool of engineers who have used it, so hiring is usually easier. Try building one real interface in each before deciding.
None of the three vendors publishes a price list as of November 2026. NextGen charges a flat annual fee per server for Mirth, iNTERFACEWARE agrees terms with each organisation, and Rhapsody does not publish prices. The Open Integration Engine is free. Ask each vendor for a written quote including renewals and every environment you need.
Its maintainers describe release 4.7.0, published on 21 October 2026, as generally available and recommended for production use, and the project is now incubating at the Eclipse Foundation. Whether it is ready for you depends on having someone to support it, in-house or under contract, since the project offers community help rather than a service level.
Partly. Google, Microsoft and AWS store and convert healthcare data well, and Microsoft converts HL7 v2 to FHIR, but none of them on its own provides the routing, queuing, retries and per-partner handling an engine gives you. Many organisations run an engine for HL7 v2 partners and a cloud FHIR service behind it.
Within the Mirth family, channels import and the work is mostly testing. Between vendor families every interface is rebuilt and re-tested, typically one to ten engineer-days each depending on complexity, plus partner coordination and a parallel run. For a large estate that is months of work, so include it in any cost comparison.
No. We are not a partner or reseller of NextGen, Rhapsody, iNTERFACEWARE, InterSystems, AWS, Microsoft or Google. Most of our interface work is on Mirth Connect and the Open Integration Engine, which is a bias you should weigh when reading this comparison. You obtain licences and trials from vendors directly.
Both engines are sold by the same company, Rhapsody. Rhapsody Integration is aimed at complex, high-volume environments, while the company presents Corepoint as a menu-driven engine suited to teams without a large IT department, available as SaaS, on your own infrastructure or as an integration platform service.
Choosing an engine, or checking a choice already made?
Tell us how many interfaces you run, which engine you use today, where it must be hosted and who will support it. We will reply with the questions that decide it and the scope of a selection assessment.






Compliance-first delivery