Skip to content

Scaffold

When a ward calls, an engineer answers.

Scaffold is the application where a hospital running Pensieve raises a support ticket and a named engineer answers it. Behind it stand the engineers who deployed your hospital, the ordered record of what the system did, and an escalation path agreed before anyone needs it.

What Scaffold is

Every ticket has a named owner and a visible clock.

Scaffold is where a hospital running Pensieve raises a support ticket, watches it, escalates it and closes it. The hospital and the Pensieve engineers who work the ticket read the same record of it: who asked, what for, and how Pensieve answered. Patient data stays in Pensieve.

Get help, from the error itself

A person looking at an error inside Pensieve presses Get help. Scaffold opens in a new tab with the ticket already composed and the error context attached. Every piece of that context is shown before it is sent, and any of it can be removed.

A named person answers

Every ticket is worked by a named engineer. The hospital claims the severity when it raises the ticket, and the ticket is worked at that severity. Where Pensieve disagrees, the engineer says so on the ticket and gives the reason.

A library anyone can read

Help articles, how-to guides, troubleshooting guides, release notes and the supported platform reference. No account is needed to read any of them.

Part of Pensieve

Scaffold is included in Pensieve at no separate charge, with no per-seat fee and no metering of how much a hospital uses it.

Who answers

The people who answer have already seen your building.

The forward-deployed engineers who produced your system inventory, installed orophin, described your hospital in Nexus Manager, rehearsed it in Cornea, and watched the parallel run are the same people in the escalation path afterwards.

Nobody asks you to describe your own configuration back to them.

Why there is no bronze, silver, or gold

The only real difference between those tiers is how long a hospital is made to wait. We will not price the removal of a delay we introduced.

What they hold

What the person answering can see.

What that team can do is bounded by the same architecture your clinicians work inside. What follows is what the person on the other end of the call actually holds.

Your configuration, as modelled

They have your sites, roles, tariffs, and forms in front of them, because they described your hospital in Nexus Manager and rehearsed it in Cornea before you went live. A proposed change can be run through that same rehearsal before it touches a live record.

Every connection, and its state

Westron shows each link to the systems you already run: which are delivering, which have stopped, and how deep the bridge outbox has grown while a downstream system was refusing messages.

One error language

The platform has one error taxonomy and one wire format. A failure raised in admissions and a failure raised in billing are shaped identically, so the first question on the call is what broke.

The ordered trace

Request tracing follows a single request through every system it touched, in sequence. What happened, and in what order, is a query here.

The audit trail as the record

Every action is written in the same transaction as the change it describes. Support starts from what the system recorded, so your staff are never asked to reassemble the night from memory.

No support back door

An engineer’s read is checked by Cirith and written to the audit trail exactly like anyone else’s. There is no privileged support account that sees past the authorization model, and every read we make is visible to you in your own logs.

Escalation

What happens when a ward calls at three in the morning

  1. It arrives on a named route

    Escalation reaches the route written into your services agreement, at the severity the hospital claims. Who is woken for each severity level is agreed in that agreement before anyone needs it.

  2. It starts from the trace

    Every request carries a trace id, and the ordered record of what the system did survives the shift change. The first hour goes on the fault itself.

  3. The scope is stated before the fix

    What is affected, what still works, and whether any record is at risk. You get the scope we can evidence from the trace and the audit trail, and you get a plain "we do not know yet" wherever the evidence stops.

  4. The fix lands in the product

    A fault is corrected on the platform release, for every hospital. Your instance is never hand-patched into a version nobody else runs, because a private build is how an institution ends up alone with software no one can maintain.

  5. You keep the account of it

    Afterwards, a written record: what failed, what caused it, and what changed so it cannot recur. The traces and audit entries behind that record are your data, and they export on the same open standards as everything else.

Confidence in hospital software is decided at the worst hour of the night, by whether the person who answers can tell you what is happening. This is the path, written down so you can hold us to it.

The path

From the ward to the release

DIAGRAMAn escalation traced left to right: a ward raises an incident on the named route; the on-call engineer opens the request trace and the platform’s error taxonomy for the failing request; Westron shows the affected connections and outbox depth; a scope statement returns to the hospital; the fix ships on the platform release for every hospital; a written account and the audit entries land back with the hospitalDIAGRAM, 16/9
One incident followed through the named route, the trace that reproduces it, the scope statement back to the hospital, the fix on the platform release, and the written account afterwards. The artifacts support reads along that path, the trace and the audit entries, are readable by the hospital too.

Refusals

What we refuse to promise.

A response time we have not agreed with you

This page publishes no response-time guarantee and no uptime percentage. Scaffold defines the severity levels; the commitment attached to each is set per hospital in the services agreement, against your deployment and your risk. A public number chosen to fit every hospital fits none of them, and it would be the first thing we quietly failed.

That the software will hold

It will fail sometimes. All software does. What is engineered here is that a failure is visible, traceable to the request that caused it, and reported in one language across the platform, and that you hear the scope from us before you hear a fix.

That we can fix your network

orophin holds a durable on-disk outbox, so machine traffic queues through an outage and delivers on reconnect. The staff web app is a cloud app and still needs connectivity. If the link into the building is down, that is your building, and we will say so.

That we can compel your other vendors

If an incumbent declines to enable an interface, that is a commercial conversation inside your institution. We will document exactly what we need, in writing, with the evidence of what is failing. We cannot run the negotiation on your behalf.

That we decide anything clinical

Who may discharge, which form is mandatory, and what an alert fires on are institutional decisions encoded as configuration, and they stay yours. What we will do is show you the consequence of a change in a Cornea rehearsal before you commit it.

Service levels

Hours and commitments are written into the services agreement.

Scaffold publishes four severity levels, from a Severity 1 that stops clinical work to a Severity 4 that is a question, each with a worked example. The hours the escalation path is staffed, the commitment attached to each level, and how often the deployment review runs are agreed per hospital in the services agreement. Ask for them in writing during the briefing, and hold us to that document.

Prevention

Before anything breaks.

The failure that costs a hospital most is a link that stopped delivering quietly, because nobody notices until a result is missing. Westron watches connection state and outbox depth continuously, and it is the same console your own team uses. Most of what keeps a hospital running is done before there is anything to call about, and none of it is billed as a change request.

Self-service

Most questions should never reach a person.

Everything below is public or already open to your own team. Reading it is faster than waiting for us, which is the point of writing it.

Deployment

The written record of how a hospital moves onto the platform: the sequence, the four deployment models, and the dependency behind every step.

Deployment

Getting started

What to do next, in order: the ways in, the sequence after you write, what we will ask you for, and the Trust Center, where the evaluation and the purchase are handled.

Getting started

Pensieve Learn

Role-scoped training. Staff learn the surface they work in, and a small group learns it deeply enough to answer their own colleagues without opening a ticket.

Pensieve Learn

Customer success

The engineers who deployed your hospital, what they stay for after go-live, and how the relationship is measured.

Customer success

The glossary

Cirith, orophin, Westron, Cornea, Durin, and the rest of the internal names on this site, defined once and used consistently.

Glossary

Westron, inside the product

Your own team watches every connection, inspects a failed message, and sees the outbox without asking us. The fastest ticket is the one you close yourself.

Westron

Sign-in

Sign-in happens on Scaffold.

There is no username and password field on this page. A credential form served from the same site as our sales copy is a phishing pattern, and a hospital should treat one with suspicion wherever it appears, including on ours. You sign in on Scaffold itself, at the address written into your services agreement, which your security team can check any link against.

A link that does not match

If you ever receive a Pensieve sign-in link that does not lead to that address, do not use it, and tell us.

Powerful. Responsible. Pensieve.

Get started