Linkiir Grid, the integration engine

The integration engine you want today.

Typed HL7 v2, FHIR, X12 and XML. Transforms in VS Code. Replay anything that breaks. Runs on your own infrastructure, with real engineers behind it.

VS Code, embedded in the product MLLP · HTTPS · SFTP · Kafka On premise, Kubernetes or VMs
How it fits

Everything in, everything out, one engine in the middle.

Grid parses what arrives, maps it with real code, validates it against a grammar you control, and replays anything that fails.

Epic MLLP · HL7 V2 Laboratory SFTP · HL7 V2 Scheduling HTTPS · FHIR LINKIIR GRID PARSE · MAP · VALIDATE Snowflake STREAM · PARQUET Availity HTTPS · X12 Analytics KAFKA · JSON SOURCES Epic MLLP Laboratory SFTP Scheduling FHIR LINKIIR GRID PARSE · MAP · VALIDATE DESTINATIONS Snowflake STREAM Availity X12 Analytics KAFKA
Inbound, as it arrives Parsed, mapped, validated Outbound, normalised
If you have run one of these before

You already know what is wrong with your engine.

Here is what we did about it.

01

Open the message, not the log file

Every message is stored parsed. Click a failure and you get the tree, the segment and the field that broke, instead of grepping for a control ID at midnight.

02

Fix it, then replay it

Change the transform and replay the affected messages yourself. No asking the sending system to resend, no apologising to a vendor, no gap in the record.

03

It is VS Code

Not a mapping wizard, not a proprietary editor from 2009. Your keybindings, your habits, real files, a debugger and a run-test button.

04

Branch it like code

Projects are Git-backed. Interfaces get branches, review and rollback, so a change to an ADT feed stops being an act of faith.

05

Prove parity before you cut over

Shadow your live traffic through Grid and diff the output against your current engine, byte for byte, before anything moves.

06

Someone actually answers

Access to engineers who have run these systems, know your estate, and are allowed to fix it rather than log it. Included, not a tier.

Speaks every protocol your integrations already speak

Epic Oracle Health · Cerner MEDITECH Expanse athenahealth eClinicalWorks Veradigm NextGen PointClickCare LabCorp Quest Diagnostics Surescripts Availity
Snowflake Databricks Palantir Foundry Azure Health Data Services AWS HealthLake Google Cloud Healthcare Apache Kafka Redpanda Salesforce Health Cloud ServiceNow DICOM · C-STORE Carequality
Browse all 197 adapters in the Integration Network
VS Code
Embedded, nothing new to learn
Ways to deploy: on prem, K8s, VM, multi-VM
Adapters in the network
Combined engine experience
One product, not a bill of materials

You get the product, and access to real engineers.

Software alone has never fixed an interface at 3 a.m. Grid ships with the people who know it. No separate services contract, no support tier to upgrade to.

Objects

Every message, a first-class object.

Grid ships a real grammar builder. Load the HL7 v2 or X12 schema, see the segments, composites and primitives, set cardinality and requirement per child, and validate before a single message moves. No regex archaeology at 2 a.m.

  • Typed access. MsgIn.PID[5][1][1] resolves to a node, not a string index you have to remember.
  • Editable grammars. Every site's ADT is a little wrong. Fork the schema instead of fighting it.
  • Parsed and raw, side by side. Flip between the tree and the JSON that produced it.
  • Matching rules. Route on message shape, not on a chain of if-statements.
grid.linkiir.com/schema
Schema Editor
HL7 / X12 message grammar builder
New Schema ⌄Import JSONSave*
Message Definition
Configure message segment references and groups.
ParsedRaw JSON HL7
ACK 5 childs · 0 matching rules ⊙ No issues
Protocol
HL7
Children
5
Rules
0
Segments
5
› Matching Rules (0) + Rule
Children
TypeRefSegmentDescriptionMinMaxReqRep
Segment ⌄1MSHMessage Header11
Segment ⌄2SFTSoftware Segment0−1
Segment ⌄3UACUser Authentication Credential01
Segment ⌄4MSAMessage Acknowledgment11
Segment ⌄5ERRError0−1
+ Segment Reference+ Segment Group

Linkiir Grid: Schema Editor

grid.linkiir.com/scripting
Scripting
Node Script Editor
ADT_A03 ⌄⚙ Debug▷ Run Test
▣ main.lua ×
require "legacy_adapter"local mapper = require "mapping" function main(Data) -- Parse incoming raw messages with the hl7.parse function. -- Returns the parsed message 'MsgIn' and message type 'MsgType' local MsgIn, MsgType = hl7.parse{ schema='demo.json', data=Data } print(MsgIn.PID[5][1][1]:nodeValue()) -- Build the outgoing HL7 message template local MsgOut = hl7.message{ schema='demo.json', name=MsgType } -- Copy all fields from MsgIn to MsgOut MsgOut:mapTree(MsgIn) -- Field-level mapping via the mapping module mapper.mapMSH(MsgIn.MSH, MsgOut.MSH) mapper.mapPID(MsgIn.PID, MsgOut.PID) -- push MsgOut to queue queue.push{ data = MsgOut:S() }end
⚡ Linkiir RuntimeLuaLn 21, Col 2822 lines · Saved

Linkiir Grid: Node Script Editor

Transforms

Write real code. Not a mapping wizard.

Transforms are Lua, in a proper editor, with a library folder, samples, a debugger and a run-test button. Copy the whole tree with mapTree, then override the six fields that are actually different.

  • Shared libraries. Write mapping.lua once, require it from every interface.
  • Run against real samples. Pin an ADT_A03 from production and iterate against it.
  • Git-backed projects. Branches, review, rollback. Interfaces are code, so treat them like it.
  • AI-native mapping. Describe the target and get a first-pass transform you can read and correct.
Runtime

Containerised. Observable. Replayable.

One console for queue depth, throughput and every service behind it. When something breaks you find the message, fix the transform and replay it, rather than restarting the engine and hoping.

grid.linkiir.com/dashboard
Dashboard
System health and overview
TE
Total Workflows
148
Active Workflows
141
Queue Backlog
12
↓ Messages In/sec
2,140
↑ Messages Out/sec
2,138
InOut last 24h
CPU Usage
31%
Memory Usage
6.4 GB
Grid Uptime
61d 4h
Workflow Health148 workflows
NameProjectStatusLast ActivityQueuedErrors
epic-adt-inRegional HealthRunning2s ago4
labcorp-resultsRegional HealthRunning1s ago0
availity-837pRCM PlatformRunning6s ago8
foundry-clinicalAnalyticsRunning3s ago0
registry-vxuPublic HealthRetrying44s ago172
nuance-mdmRegional HealthRunning11s ago1
System Status4 services
RuntimeOnline
KafkaOnline
Notification EngineOnline
Archiver (archiver-1)Online

Linkiir Grid: Dashboard

Message-level durability

Every message is queued, acknowledged and archived before the next hop. Nothing depends on a process staying alive.

Replay, not restart

Select a window, a workflow or a single message. Fix the transform, replay, watch it clear.

On prem, Kubernetes or VMs

One artifact, four ways to run it: on premise, on Kubernetes, on a single VM, or across multiple VM deployments. High availability with automatic failover when you need it, a single node when you don't.

Coming soon · 3 months

Linkiir Cloud

Fully managed Grid, hosted and operated by us. Same engine, same engineers, none of the infrastructure.

Migration

Replace the engine you use. Only newer.

We came out of the team that built one of these. We know what your Rhapsody routes look like, why your Cloverleaf TCL is untouchable, and which Mirth channel nobody will admit to owning.

RhapsodyMirth ConnectIguana CloverleafCorepointBizTalk Ensemble / HealthShareHome-grown

First interface live in an hour

Download the package, point a source at your test feed, get an HTTP endpoint you can curl.

Parity proven in two weeks

We shadow your live traffic through Grid and diff the output byte-for-byte against the incumbent. You see the report before you cut over.

Cut over interface by interface

No big-bang weekend. Both engines run until the last route moves, and rollback is a routing change.

“Bring us your worst interface. The one with the 400-line script, the vendor who won't answer, and the field that means something different on Tuesdays. That's the one we want to see.”

LK The Linkiir engineering team
Integration engineers, not account managers
What we ask for
  • A sample of real messages, de-identified
  • The current mapping, however it's written
  • Two hours with whoever knows the quirks

You get back a working Grid workflow and a written parity report. No charge, no obligation.

LNK-4821 · Jira Service Management
DK
DanaAcme Health

registry-vxu is retrying with ACK^AE, “VXU-3 invalid”. Nothing changed on our side.

RS
Ravilinkiir

Looking. Queue is holding, nothing lost. The registry started rejecting CVX 213 earlier this morning, so it looks like their validation rather than your mapping.

RS
Ravilinkiir

Confirmed with them. Mapping added for the new code. Replay now, or wait for your morning?

Platform services

An engineer, every time.

No tier-1 gatekeeping. No “have you tried restarting the engine.” You reach integration engineers who know your interfaces and can act on the estate, rather than file a ticket about it.

Most of what people want from support is not a response-time number on a contract. It is someone who recognises the problem, and is allowed to fix it.

Security & governance

Built to pass the questionnaire you're about to send us.

PHI-grade controls in the product, not in a roadmap deck, starting with the fact that we don't need to see your patient data to support you.

We never see PHI

We diagnose from structure, metadata and error codes, never patient data. We don't ask for PHI, and we advise you not to send it, even to us.

Encrypted end to end

TLS 1.2+ in transit, AES-256 at rest, customer-managed keys on request. Message archives encrypted per tenant.

RBAC, MFA, SSO

Role-based access down to the project, MFA enforced, SAML SSO and SCIM provisioning. Included, not gated behind a tier.

HIPAA · PIPEDA · PHIPA

Aligned controls, BAA-ready, Canadian and US data residency. Immutable audit log of every access and action.

Integration Network

197 adapters. Pre-tested, documented, versioned.

EHRs, labs, PACS, pharmacy, clearinghouses, registries, HIEs, data platforms and the message brokers in between. Every adapter ships with a sample project, a schema and a support owner.

Adapters
EHR & EMR
Message formats
Source
Epic ADT
Transform
Map & enrich
Destination
Foundry
grid.linkiir.com/logs

Live message log: every hop, every ack

Tell us about your interfaces

Bring us your worst interface.

Thirty minutes with an engineer who has seen your exact problem before. No SDR round-trip, no discovery deck. We'll open the message and look at it with you.