Skip to content

Exporting to intu-dev and running it yourself

Exporting a project as a portable intu-dev project, running it on the open-source engine, and where to go for engine-level documentation.

Last updated August 12, 2026

On this page

Every IntuCloud project is an intu-dev project. Exporting gives you the files; the open-source engine runs them.

What you get

An export contains:

  • channel.yaml for each channel
  • The TypeScript stage files for every enabled stage
  • Project configuration and test data fixtures

Secrets are not included — they are references, and you supply the values in your own environment.

Running it

With the engine installed:

intu validate
intu serve

intu serve is the same command in every environment. There is no separate production runtime and no application server to install — the engine is a Go binary of roughly 20 MB plus a Node.js worker pool for your TypeScript.

What you take on

Self-hosting means owning the parts IntuCloud was handling:

  • Infrastructure — where the engine runs, and how it is kept running.
  • Networking — listener addresses, TLS certificates, allowlists.
  • Storage — pick a backend: memory, Postgres, MySQL, MSSQL, SQLite or S3.
  • Observability — the engine emits OpenTelemetry traces and Prometheus metrics; you point them somewhere.
  • Availability — Redis-based clustering is included in the open source, but you run the Redis.
  • Compliance — the data is inside your boundary, which is often exactly the point.

What you gain

  • Every connector. Cloud-beta gaps do not apply. DICOM listeners, database and file pollers, Kafka consumers, FHIR poll and subscription all work.
  • No metering. Message volume is bounded by your hardware, not a plan.
  • No data leaves your network.

Keeping support

Running the engine yourself does not mean being on your own with it. Enterprise support subscriptions provide a named technical contact, response-time SLAs, assisted upgrades and priority fixes for self-hosted deployments.

Going the other direction

An intu-dev project imports back into IntuCloud. The decision is reversible in both directions, which is a reasonable thing to verify before you depend on it.

Engine documentation

CLI reference, channel YAML schema, clustering topology, storage backend configuration and deployment guides live in the intu-dev docs. The source is on GitHub under MPL-2.0.