Home / Products / OpenLI TIE Migration
Healthcare & Operations · production

Manage any NHS Trust Integration Engine migration in plain English.

OpenLI TIE Migration turns highly specialised integration engineering into a governed, AI-assisted programme any NHS healthcare IT team can run themselves. Migrate Rhapsody or Mirth Connect estates to InterSystems IRIS Health Connect through a 5-stage lifecycle with phase-gate approvals, audit trails on every action, and plug-and-play AI runners.

Built on the OpenLI Codex foundation. The same governance, RBAC, audit trail and dual-runner architecture as every other OpenLI product — one security review covers your entire OpenLI portfolio.

Open TIE Migration Portal View on GitHub Talk to the team
v0.8.12 production 5-stage lifecycle Plug-and-play runners Whitelabel-ready tm.openli.ai
5Lifecycle stages with phase-gate approvals
70Jinja2 prompt templates across all stages
9Automated QA checks on ObjectScript output
3AI runners (Mock free, Codex, Claude)

A migration programme, in plain English.

NHS Trust Integration Engine migrations have historically been long, expensive and dependent on a small pool of specialist contractors. OpenLI TIE Migration changes that. Internal NHS digital teams describe what they want in English; the runtime drafts the design, generates the ObjectScript, runs QA, plans the deployment and produces the rollback artefacts — under human approval at every gate.

5-stage lifecycle with phase gates

Onboarding → Landscape Analysis → IRIS Design → Migration Planning → Implementation. Each stage requires explicit approval before the next can begin. Gates are defined per project, not hard-coded.

Plug-and-play AI runners

Choose between Mock (free, deterministic), OpenAI Codex (paid), and Claude Agent SDK (paid) per task. Switch runners mid-project. The Mock runner lets evaluators run the full workflow at zero cost before committing budget.

Cost dashboard

Token tracking by input, output and cached. USD cost breakdown per model per stage. Per-Trust budget envelopes. No more migration projects with surprise AI bills.

Capabilities (v0.8.12, in production today)

The capabilities below ship today. They are the actual product running at tm.openli.ai.

Interface Catalogue auto-extraction

Auto-extract the source interface catalogue from Rhapsody RLC or Mirth Connect XML. The runtime understands both formats and produces a normalised catalogue ready for analysis.

5-stage lifecycle

Onboarding, Landscape Analysis, IRIS Design, Migration Planning, Implementation. Each stage has its own prompt templates, its own QA checks, and its own approval gate.

QA engine (9 automated checks)

Generated ObjectScript is checked for syntax, naming conventions, error handling, transport configuration, namespace correctness, business host conformance and more — before a human reviewer ever sees it.

IRIS XML export

Standards-compliant InterSystems IRIS XML output, importable directly via $SYSTEM.OBJ.Load(). Drop the export into your IRIS instance and the productions are ready to test.

70 prompt templates

A library of 70 Jinja2 prompt templates covering all 5 stages plus AI skills (nhs-tie-migration, iris-objectscript, hl7-mapping, tie-quality-reviewer, tie-reverse-engineering).

Cost dashboard

Token tracking with input, output and cached counts. USD breakdown per stage and per model. Budget envelopes per project. Surprise-free AI cost management.

Enterprise RBAC

5 role tiers: super_admin, org_admin, project_admin, editor, viewer. Each Trust gets its own isolated workspace with its own audit log.

Reference benchmark

A UK NHS Foundation Trust migration of 44 interfaces, 60 ObjectScript classes and 7 HL7 test fixtures provides the reference benchmark for the platform.

Built on OpenLI Codex

Same dual-runner architecture, same governance, same audit framework as the rest of the OpenLI portfolio. One security review for all OpenLI products.

How a migration runs

1. Onboarding

Configure the project, the source engine type (Rhapsody or Mirth), and the target IRIS environment. Approve the project to enter Landscape Analysis.

2. Landscape Analysis

Auto-extract the source interface catalogue from the engine’s native config files. The runtime classifies and clusters interfaces by transport, message type and downstream system.

3. IRIS Design

For each interface, the runtime drafts the IRIS Health Connect design: namespace, production, business hosts, transformations. Reviewers approve each design before code generation.

4. Migration Planning

Sequence the migration into deployable waves. Identify dependencies, risk hotspots and rollback points. Generate a migration plan reviewable by the integration lead and clinical safety reviewer.

5. Implementation

Generate ObjectScript, run the QA engine (9 checks), produce the IRIS XML export, and ship to the deployment pipeline. Audit trail records every step.

Try the live evaluation portal

The OpenLI TIE Migration evaluation portal is hosted at tm.openli.ai. Request access to evaluate the platform with the Mock runner (free) or with your own OpenAI / Anthropic API keys.

Open tm.openli.ai Request evaluation access
FOUNDATION

Built on OpenLI Codex

The dual-runner architecture (Claude Agent SDK + OpenAI Codex SDK), the prompt template library, the workspace registry, the RBAC engine and the audit trail all come from the OpenLI Codex foundation. Adopt OpenLI TIE Migration and you adopt the foundation that powers HIE, IRIS CoPilot, IMX Monitor, OpenLI Data Migration and OpenCT.

Read the OpenLI Codex foundation story See the product family tree