A TECHEMV PRODUCTThe HL7 and FHIR integration engine with generative AI.
Describe what you need in natural language and Aurora builds the complete channel — source, transformations, destinations — production-ready.
- channel
- ADT_A01_to_FHIR_Patient
- source
- MLLP :2575
- map
- PID → FHIR Patient
- send
- REST POST /fhir/Patient
- ack
- ACK / NACK configured
- 12+
- native connector types
- 0
- engine restarts to update a channel, with hot-reload
- 12 ms
- median ACK latency at 100 msg/s
- 5
- UI languages
Aurora in 74 seconds. The full demo takes you from a natural-language description to a channel in production.
01 · The AI assistant
Describe the channel. Aurora builds it.
The AI assistant is built into the engine: it creates channels from scratch from a natural-language description or a specification, modifies existing ones and explains how Aurora works.
Describe
In plain language or by pasting a specification: "ADT A01 over MLLP on port 2575 to FHIR Patient via REST".
Aurora generates
Source, transformations, destinations, the HL7 → FHIR mapping and ACK and NACK handling: the complete channel, not a suggestion.
Modifies and explains
The same assistant modifies existing channels and answers questions about how Aurora works.
Into production
The channel is a YAML configuration: if it is valid the engine accepts it and, with hot-reload, it goes live without restarts.
Models: OpenAI, Anthropic or Ollama running locally. With Ollama no data leaves your network.
02 · The problem
Legacy integration engines hold back innovation.
Traditional platforms slow teams down, expose risks and drain budgets.
Interfaces from another era
Java desktop clients and GUIs that take weeks of training.*
Inadequate security
Plain-text credentials in configuration files, no multi-factor authentication, no user roles: limits that are hard to square with GDPR.*
Exploding budgets
Five- or six-figure licences, consultants for every change, total vendor lock-in.
* Mirth Connect open source, version 4.5.2, the last one released under an open-source licence, in its default configuration. Multi-factor authentication and user roles are paid extensions from NextGen Healthcare.
03 · Features
Simplified integration, accelerated results.
Built from scratch with Kotlin, GraalJS and generative AI. Enterprise power, consumer simplicity.
Generative AI, not just assistive
Aurora generates complete channels from natural language. It works with OpenAI, Anthropic and Ollama for fully local AI: no data leaves your network.
A professional editor in the browser
Monaco Editor, the one inside VS Code, built in. JavaScript with GraalJS and Groovy, light and dark themes.
Uncompromising security
AES-256-GCM with a machine-identity key, four-level RBAC, TOTP 2FA, audit trail and account lockout. GDPR by design.
Hot-reload and container-ready
Every HTTP channel runs on its own Netty server: edit and deploy without restarts. Multi-stage Docker image and a Helm chart for Kubernetes.
Tracing, DLQ and monitoring
Large OBX segments with base64 PDFs and images shown with per-field truncation. Inline HL7 validation, dead letter queue, compressed NDJSON archive.
AI-assisted migration from Mirth
The migration tool analyses Mirth channels, converts scripts and generates the Aurora configuration. An alternative to Mirth's commercial licensing from version 4.6.
Connectors and standards
04 · Comparison
Aurora and the most widely used engines, on published facts.
The grid holds only what each vendor publishes, with sources and dates. Where a page does not state something, we say so instead of guessing.
| Criterion | Aurora | Mirth Connect | Rhapsody · Corepoint | InterSystems Health Connect | Infor Cloverleaf | Iguana X |
|---|---|---|---|---|---|---|
| Transformation languages | JavaScript (GraalJS) and Groovy | JavaScript (Rhino) and Groovy | Low-code tools and scripting components | ObjectScript, Python, Java, .NET, SQL or graphical editors | Tcl, Python and JavaScript | Lua |
| Generative AI in the product | Creates and modifies channels from a description and explains how the engine works; models can run locally | Not stated on the licence tiers page | Axon, an embedded AI agent (since March 2026): guides and automates steps of integration workflows | Tools to build generative AI applications, such as vector search; production generation not stated | Not stated on the pages consulted | A public ChatGPT GPT that writes Lua code for components, outside the product |
| Benchmark with Aurora | — | Yes: seven eight-hour runs against 4.5.2, report published | Not tested | Not tested | Not tested | Not tested |
Sources consulted on 29 September 2026: NextGen Healthcare, Mirth Connect licence tiers page and announcement of 19 March 2025; technical guides on Mirth Connect (mirth.support, April 2026); Rhapsody, Axon announcement of 4 March 2026 and information page of July 2026; InterSystems, IRIS for Health and Health Connect 2026.2 documentation; Infor, Cloverleaf page and product listing; iNTERFACEWARE, Iguana X documentation and blog of 1 May 2026. All trademarks belong to their respective owners. Something out of date? Tell us and we will correct it.
05 · Benchmark
Measured, not claimed.
Seven eight-hour runs against Mirth Connect 4.5.2 on the same hardware, with the same traffic and the same simulator: synthetic HL7 messages at 100 per second and a real production channel with remote signing, document repository upload and ward notification.
The other engines were not tested. Method, charts and full data are in the report. Download the full report (PDF) →
06 · Where it runs
Built for healthcare. Ready for any industry.
Healthcare
HL7, FHIR, PACS, LIS, RIS. Hospital systems, laboratories, radiology. GDPR and audit trail.
Laboratories
OBX and OBR, LIMS. Multiple databases: PostgreSQL, SQL Server, Oracle.
Enterprise
APIs, ERP and CRM, B2B, Kafka and RabbitMQ. A modern ESB in place of legacy middleware.
Evaluating an integration engine, or moving off Mirth Connect?
Bring your projects: the demo runs on your own cases.