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.
Grid parses what arrives, maps it with real code, validates it against a grammar you control, and replays anything that fails.
Here is what we did about it.
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.
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.
Not a mapping wizard, not a proprietary editor from 2009. Your keybindings, your habits, real files, a debugger and a run-test button.
Projects are Git-backed. Interfaces get branches, review and rollback, so a change to an ADT feed stops being an act of faith.
Shadow your live traffic through Grid and diff the output against your current engine, byte for byte, before anything moves.
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
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.
Typed objects for HL7 v2, FHIR, X12 and XML. Lua transforms in an embedded VS Code, message-level durability and replay. On premise, Kubernetes, a single VM or multiple VMs.
Platform overview IncludedYou get the product and access to real engineers. Incidents, break-fix, new interface builds and documentation, on Grid and on the estate you already have.
Platform services Agent · March 2027Duplicate ADTs, patient-match drift, rejection clusters and quiet endpoints, detected, explained and fixed through Grid or your existing engine. Every action a signed record. Launching March 2027.
Preview what Agent will doGrid 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.
| Type | Ref | Segment | Description | Min | Max | Req | Rep |
|---|---|---|---|---|---|---|---|
| Segment ⌄ | 1 | MSH | Message Header | 1 | 1 | ✓ | |
| Segment ⌄ | 2 | SFT | Software Segment | 0 | −1 | ✓ | |
| Segment ⌄ | 3 | UAC | User Authentication Credential | 0 | 1 | ||
| Segment ⌄ | 4 | MSA | Message Acknowledgment | 1 | 1 | ✓ | |
| Segment ⌄ | 5 | ERR | Error | 0 | −1 | ✓ |
Linkiir Grid: Schema Editor
Linkiir Grid: Node Script Editor
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.
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.
| Name | Project | Status | Last Activity | Queued | Errors |
|---|---|---|---|---|---|
| epic-adt-in | Regional Health | Running | 2s ago | 4 | – |
| labcorp-results | Regional Health | Running | 1s ago | 0 | – |
| availity-837p | RCM Platform | Running | 6s ago | 8 | – |
| foundry-clinical | Analytics | Running | 3s ago | 0 | – |
| registry-vxu | Public Health | Retrying | 44s ago | 17 | 2 |
| nuance-mdm | Regional Health | Running | 11s ago | 1 | – |
Linkiir Grid: Dashboard
Every message is queued, acknowledged and archived before the next hop. Nothing depends on a process staying alive.
Select a window, a workflow or a single message. Fix the transform, replay, watch it clear.
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.
Fully managed Grid, hosted and operated by us. Same engine, same engineers, none of the infrastructure.
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.
Download the package, point a source at your test feed, get an HTTP endpoint you can curl.
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.
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.”
You get back a working Grid workflow and a written parity report. No charge, no obligation.
registry-vxu is retrying with ACK^AE, “VXU-3 invalid”. Nothing changed on our side.
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.
Confirmed with them. Mapping added for the new code. Replay now, or wait for your morning?
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.
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 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.
TLS 1.2+ in transit, AES-256 at rest, customer-managed keys on request. Message archives encrypted per tenant.
Role-based access down to the project, MFA enforced, SAML SSO and SCIM provisioning. Included, not gated behind a tier.
Aligned controls, BAA-ready, Canadian and US data residency. Immutable audit log of every access and action.
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.
Live message log: every hop, every ack
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.