Shop-floor execution for engine MRO
Torque is the shop-floor layer for engine MRO. Measurements, photos and signatures are captured at the bench instead of retyped at a desk — and the work package assembles itself out of the manual pages you worked from. Your Quantum, Corridor or QuickBooks stays exactly where it is.
Compressor turbine blade — chord
CT shroud segment — radial clearance
Hot section — visual, gas generator case
Torque uses the vocabulary of the engine shop: travelers, ATA chapters, findings, life-limited parts and release to service.
The gap
Every shop has a version of the same handoff, and every shop loses something in it. Not because the people are careless — because the tool that holds the record was never designed to be held in a gloved hand next to a hot section.
Two links in the middle carry no enforcement, no timestamp and no author. Everything a regulator cares about is decided there.
A dimension is read at the bench, pencilled onto a smudged card, then typed into a system at a desk an hour later. Every transcription is a chance to transpose a digit — and on a life-limited part, a transposed digit is not a typo, it is an event.
Disassembly, inspection and build-up are tracked. The paperwork that releases the engine is assembled at the end, by hand, by whoever is free — which is why the last days of a turnaround are so often spent chasing a signature rather than turning a wrench.
An auditor asks which revision of the manual governed a dimensional check performed nine months ago. If the answer is "the current one", that is not an answer. The limits in force at the moment of execution have to be part of the record, not reconstructed from it.
On the record
“approximately 50% of this time was spent completing paperwork”
On where MRO engineers' hours actually go.
“a non-RII inspector cannot sign an RII required block”
On what an electronic task card has to enforce.
“manual, redundant data entry performed hours after work is completed”
An incumbent MRO platform, describing the problem on its own blog.
“audit paperwork in real time”
On what changed once the record was written at the aircraft.
Quoted for context, not endorsement. None of these organisations is affiliated with Torque.
ENGINE WORK CARD
PT6A-38 · OVERHAUL · SHEET 3 OF 11
| # | ATA | ITEM | MEAS. | LIMIT | MEC |
|---|---|---|---|---|---|
| 1 | 72-50 | CT blade — chord | 1.874 | 1.860 / 1.890 | MD |
| 2 | 72-30 | Shroud — radial clr. | 0.0213 | 0.0150 / 0.0210 | MD |
| 3 | 72-00 | GG case — visual | OK | Photo req. | MD |
| 4 | 73-10 | FCU — drive torque | 42.0 | 40.0 / 45.0 | MD |
Work order
OT 5366
CT blade — chord
Shroud — radial clr.
GG case — visual
FCU — drive torque
The document you already trust
A ruled header carrying the engine's identity. Numbered items down the page. A measurement written beside its limit. A stamped block where a named inspector accepted the work.
Nothing about it is wrong
It is dense, it is unambiguous, and nobody has ever been confused about which part of it matters. The problem was never the document. The problem is that paper cannot check anything.
So keep the document
Torque is that card, executed. Same ruled sections, same numbered items, same stamped block — except the limit is checked as you type, and the stamp is issued by the server under a named account.
On the floor
This is the real sheet, running here. Tap a task, enter a measurement and watch what happens — the second one is deliberately out of limit, so try it.
Minimum and maximum come from the revision assigned to that engine model. The technician never leaves the task to look them up, and never works from a remembered number.
The sheet accepts it and raises a finding against it. A tool that refuses a bad measurement is a tool that teaches people to write down a good one.
The completion stores the technical-data reference, its revision and the limits that were in force at that moment. Nine months later the record still knows what it was judged against.
Work order
OT 5366
PT6A-38 · S/N PCE-RB0421
CT blade — chord at root
CT shroud — radial clearance
FCU — drive shaft torque
Finish task
Task
Interactive · the numbers and limits are real PT6A-shaped values
The gate
On paper, a stamp is a claim. Here it is a record: a named account, a server timestamp, and a defined thing it applies to. That is the whole difference, and it is the part an auditor cares about.
Inspector review
OT 5366
| ATA | Item | MEC | INSP |
|---|---|---|---|
| 72-50 | CT blade — chord at root | M. Duarte | R. Salas |
| 72-30 | CT shroud — radial clearance | M. Duarte | Finding open |
| 73-10 | FCU — drive shaft torque | J. Peralta | R. Salas |
| 72-00 | GG case — visual | M. Duarte | R. Salas |
Illustrative render of the review surface. The finding on ATA 72-30 is the out-of-limit clearance from the sheet above — it holds the order open until it is resolved against a conforming execution.
The dossier
Not a summary of the job. The manual pages themselves, with your answers printed where the blanks were, every signature attributed, and a checksum proving the source was the one you activated.
OT 5366 · Hot section inspection
PT6A-38 · S/N PCE-RB0421 · Overhaul
Status: reviewed by inspector · version 7
Execution record: M. Duarte · 2026-03-14T14:07:52Z
Inspection record: R. Salas · 2026-03-14T16:02:11Z
Paper signature · Technician: ______________ Inspector: ______________
PT6A-38 Maintenance Manual · rev. 34
sha-256 4990e88caa792cead823cf41c5e86a0a43c03fa97db87176ef9b601c2e791ec0
72-30 · p. 14 · Radial clearance, CT shroud segment
P/N 3036925 · S/N PCE-RB0421-14
MEC: M. Duarte · 14:07:52Z INSP: —
Measurement: 0.0213 in — above limit
Requirements at execution: PT6A-38 MM rev. 34 · limits 0.0150 / 0.0210
72-50 · p. 9 · Chord at root, CT blade
MEC: M. Duarte · 14:02:10Z INSP: R. Salas · 16:01:44Z
Measurement: 1.874 in — within limit
Annex · values exceeding the space on the original
1 · 72-30 remarks — see entry above
Illustrative render of an exported package. Structure, field order and the checksum line follow the product's export.
The referenced manual pages are pulled from the stored source PDF and placed in the package. Any form fields already on them are cleared and flattened first, so a package can never inherit a value from a previous order. The recorded answers are then drawn at the original field coordinates.
The stored original is fetched and its SHA-256 compared against the hash recorded when the revision was activated. If they differ, the export aborts rather than producing a document nobody can vouch for. The hash is printed in the package.
Each execution prints its own requirements line: the technical-data reference, its revision, and the minimum and maximum that applied at the moment the work was signed.
An out-of-limit measurement is printed, and so is its resolution — who accepted it, when, and against which conforming execution. The record shows the rework, not a clean second attempt.
A value too long for its box on the original page is printed in a numbered annex instead of being clipped. The reader is told where to look.
Content, approval and history are read in a single stable snapshot, so the body of the package and the approval stamped on it can never describe two different moments.
Under the hood
These are catalogue counts — extracted candidates and mapped fields, not technically approved content. Every revision is reviewed and activated by an administrator in your shop before a technician can work from it.
Where it fits
Ripping out the system that holds your inventory and your accounting is a year of your life. Torque does not ask for it. It takes the one part those systems were never shaped for — the hours between opening an engine and signing it off — and leaves everything else alone.
Stays where it is
Quantum · CORRIDOR · QuickBooks · your records system
Moves to Torque
The shop floor, and the document it produces
A note on the incumbents, since nobody else will say it: Quantum Control and CORRIDOR are both CAMP Systems products, and CAMP has announced it is bringing them into a single new platform. That is a fine reason to keep your system of record and a poor reason to wait before fixing the bench.
Pricing
Per-technician pricing is how a shop-floor tool ends up living on one laptop in the office. If it costs you more every time you hand it to another mechanic, you will hand it to fewer mechanics — and then it does not work. Torque is banded by engines through the shop per year. Put it on every phone.
One shop. The execution layer, end to end.
USD 850 – 1,500
USD / month · billed annually
Your own manual revisions, mapped and working.
USD 1,500 – 2,500
USD / month · billed annually
More than one shop, isolated from each other.
USD 3,500 – 6,000
USD / month · billed annually
One-time
These are bands, not a checkout. Where you land inside one depends on engines per year, how many models you need mapped, and how much of your documentation is already digital. The first shops get design-partner terms in exchange for being wrong with us in public — we will put that in writing before you pay anything.
Questions
No software is, and any vendor who says otherwise is telling you something that cannot be true. The FAA and EASA certificate organisations and people — not tools. Torque is built for Part 145 and EASA Part-145 engine shops, and its signature and recordkeeping behaviour is designed against the standards in FAA AC 120-78B, Electronic Signatures, Electronic Recordkeeping and Electronic Manuals, issued 11 December 2024. Your accountable manager and your quality system still own compliance. We just make the evidence easier to produce and harder to lose.
No. Those are issued by your certificated staff, under your certificate. Torque assembles and holds the evidence behind them — the executions, the measurements against their limits, the evidence files and the signatures — and exports it as one package. For the avoidance of doubt: the EASA Form 1 lives in Appendix II to Annex I (Part-M) of Regulation (EU) No 1321/2014, and an 8130-3 is never an approval to install on a given engine. Neither is something a piece of software can hand you.
Yes, and this is the part worth testing in a demo. The mechanic signature and the inspector signature are separate authenticated actions. A mechanic cannot fill the inspector slot. Only the inspector assigned to that order can, and only after the mechanic signature exists. It is enforced in the database, not by hiding a button — a request forged past the interface is refused and the order stays in review. What we do not claim is to make you "RII compliant": your air carrier customer designates which items are RII, and that decision is theirs.
No, and we would rather you did not. Torque is an execution layer. Inventory, purchasing, accounting and your engine records stay in the system that already holds them. The whole reason this product exists as a separate thing is that replacing your system of record is a year of your life and this problem is not worth a year of your life.
Not for work orders. Torque needs a connection at the moment you save. We know that is a real constraint — poor shop-floor coverage is the most common reason digital task cards get abandoned, and large MROs have said so publicly. We would rather you walk your bays with a phone and find the dead spots before you buy than have us claim a queue that does not exist.
The catalogue that is already imported and mapped is PT6A: 185 manual revisions, 5,380 extracted items and 30,568 fields mapped to their coordinates on the original page. Those are extraction counts — an administrator in your shop reviews and activates a revision before any technician can work from it. Other models are a mapping exercise, which is what the one-time catalogue fee covers.
We are not going to print a number we have not earned. We have not done this with enough shops to promise a timeline, and a fabricated one would be found out in week two. What we will do on the call is walk your actual workflow and tell you honestly which parts are configuration and which parts are work.
An aeronautical technician, inside a working turboprop engine shop, because the tools on offer were priced for an airline and shaped for a desk. The shop runs on it. That is the entire origin story, and it is also why the vocabulary is right: nobody had to explain what a hot section is, or why a transposed digit on a life-limited part is not a typo.
Request access
Independent, turbine or turboprop, doing your own teardowns. If you already have a system of record you are broadly happy with, that is a point in your favour, not against.
Not a slide deck. You describe how an engine moves through your shop and we tell you which parts Torque already does, which parts are configuration, and which parts it does not do.
We map enough of your documentation to run a single job — bench to signed package — on your own engine model, with your own people. If it does not survive that, there is nothing to discuss.
Price, scope and what happens to your data if this does not work out. Agreed before anything is paid.