Skip to content
IntuCloud is built on the open-source intu-dev engine — star it on GitHub

Open Source

The open-source engine at the heart of IntuCloud.

intu-dev is a Git-native, AI-friendly healthcare interoperability framework licensed under MPL-2.0 — Integration as Code for Healthcare. It is the engine that runs every IntuCloud channel, and it is yours to run too.

Star on GitHub Read the docs

What it is

A production engine in a single command.

The engine is a Go binary of roughly 20 MB paired with a Node.js worker pool that executes your TypeScript stages. A channel is a channel.yaml file plus a folder of TypeScript transformer files — so an interface lives in version control alongside the rest of your code, diffs like code, and can be generated or reviewed by an assistant like code.

There is no application server to install, no GUI-plus-database to back up, and no separate runtime for production. intu serve is the same command on a laptop and in a cluster.

View the source intu.dev docs

License

MPL-2.0

A real open-source license, not a source-available one with a usage clause.

Footprint

~20 MB binary

No JVM. Designed for sub-millisecond transforms once the worker pool is warm.

Connectors

27 types

14 sources and 13 destinations — all of them available when you self-host.

Breadth

Everything the engine speaks.

Self-hosted deployments have no cloud-beta gaps: every connector and standard below is available today.

HL7 v2

MLLP and raw TCP framing, ACK/NACK, segment-level access in TypeScript.

FHIR R4 / R4B

Server listener, scheduled poll, and R4B subscriptions.

CDA / CCDA

Clinical document parsing and generation, with a document viewer.

X12

EDI transaction sets for eligibility, claims and remittance workflows.

DICOM

C-STORE SCP and SCU for imaging instance transfer.

IHE ITI

XDS.b ITI-41/42/43 document sharing, PIX ITI-8/9, PDQ ITI-21/22.

Direct

Direct secure messaging for provider-to-provider exchange.

SOAP

SOAP 1.1/1.2 services on both the receiving and sending side.

Raw / JSON / XML / CSV / binary

Everything else that shows up on a hospital network.

Sources (14)

  • HTTP
  • TCP (raw / MLLP)
  • FHIR
  • SOAP
  • IHE
  • File (local / FTP / S3 / SMB)
  • SFTP
  • Kafka
  • Database
  • Email (IMAP / POP3)
  • DICOM
  • FHIR Poll
  • FHIR Subscription
  • Channel

Destinations (13)

  • HTTP
  • FHIR
  • TCP
  • Kafka
  • Database
  • File
  • SFTP
  • SMTP
  • Direct
  • Log
  • DICOM
  • JMS (ActiveMQ)
  • Channel

In the IntuCloud beta some listener types are not exposed yet — see the cloud availability column for the honest list.

Built for production

Not a demo that grew up.

Retries and dead letters

Constant, linear and exponential retry policies with jitter, and a dead-letter queue for what still will not go through.

Clustering, included

Redis-based clustering ships in the open source. Running more than one node for availability does not require a commercial license.

OpenTelemetry and Prometheus

Traces and metrics out of the box, so the engine fits the observability stack you already run.

Secrets from where you keep them

HashiCorp Vault, AWS and GCP secret managers, resolved at runtime and never written into logs or traces.

Storage backends

Memory, Postgres, MySQL, MSSQL, SQLite and S3 — pick what your operations team already runs.

Hot reload

Change a channel file and the engine picks it up without a restart.

Dashboard auth

The bundled dashboard authenticates against OIDC or LDAP, so self-hosted deployments fit existing identity infrastructure.

A ~20 MB Go binary

No JVM, no application server. Designed for sub-millisecond transforms, and small enough to ship anywhere a container runs.

Sub-millisecond transform figures describe what the architecture is designed for — a Go core with a warm TypeScript worker pool and no JVM — not results from a published benchmark.

CLI

The whole engine from a terminal.

  • Command
    intu init
    What it does
    Scaffold a new project with a channel and stage files.
  • Command
    intu serve
    What it does
    Run channels locally or in production — the same command in both.
  • Command
    intu validate
    What it does
    Check channel YAML and stage types before deploying anything.
  • Command
    intu build
    What it does
    Compile the project into a deployable artifact.
  • Command
    intu channel
    What it does
    List, describe, clone, export and import individual channels.
  • Command
    intu deploy
    What it does
    Push a built project to a running engine.
  • Command
    intu stats
    What it does
    Message counts, error rates and throughput for running channels.
  • Command
    intu reprocess
    What it does
    Replay messages after a downstream outage.
  • Command
    intu prune
    What it does
    Trim stored message history according to your retention policy.
  • Command
    intu import mirth
    What it does
    Convert Mirth Connect channel XML into an intu-dev project.
  • Command
    intu dashboard
    What it does
    Serve the bundled web dashboard.

Open core

Run it three ways.

The engine stays free. What you can buy is somebody else running it, or somebody else being responsible when it breaks.

Self-host, free forever

Download the binary, write channels, run them. Every connector, no seat count, no license key, no phone-home.

Read the docs

Self-host with enterprise support

Keep PHI inside your own network and put the engineers who build the engine on call, under a response-time SLA.

Enterprise support

Let IntuCloud run it

The same binary, pinned by digest from sandbox to production, with the designer, environments, monitoring and PHI under BAA on top.

See pricing

Run the engine, however you want to run it.

Self-host it free, buy support for it, or let IntuCloud operate it for you.

Read the docs See pricing