Digitize your vintage cutter without touching the panel

The problem

The most expensive asset in the cutting room leaves no record.

the cutter is the most expensive asset in the cutting room and the one that determines real fabric yield. Today nobody correlates cutting parameters with fabric waste because the panel's data dies every time the screen rotates.

• The automatic spreader-cutter lays and cuts the fabric lay, and it decides the real fabric yield. In most workwear plants it is over a decade old. It still cuts well, which is exactly why nobody can justify replacing it.
• Its panel shows the data that matter — cutting speed, pressure, lay height, events — and that data vanishes every time the screen rotates. The operator sees it for a few seconds; the plant never sees it at all.
• So nobody can correlate cutting parameters with fabric waste: the marker efficiency is declared, the real yield is argued about afterwards. On coverall and work shirt fabrics, a point of yield lost per order adds up to rolls a month.
• Replacing the machine just to get the data out costs hundreds of thousands of dollars, so the question stays unanswered order after order. Meanwhile the same waste repeats on every coverall and work shirt order that goes through the machine.

How it fits the IRIS system

Connect in photo-on-panel mode — the cutter stays exactly what it is.

Connect photo-on-panel mode. A fixed industrial camera reads the screen every few seconds; OCR with an anchored LLM interprets every frame and structures the cutting history by production order, cross-referenced against the declared yield percentage.

It does not integrate with the cutter's controller, opens no ports and installs nothing on the machine. A camera reads what the operator already sees, and the plant gets every cut's history structured by production order. The machine that decides how much fabric you consume finally explains itself.

See the full IRIS architecture →

Before and after

Today's cutter versus the cutter being read

Aspect	Today	With iLEAN Connect
Cutting parameters	On screen, then gone	Digitized per cut
Link to the production order	None	Structured by order
Declared vs. real yield	Argued about	Cross-referenced per order
Root cause of fabric waste	Impossible to trace	Analysis in minutes
The cutter's controller	—	Untouched
Replacement CAPEX in workwear sewing	In next year's plan	Indefinitely deferrable

zero correlation between cutting parameters and fabric waste → root-cause yield analysis per order in minutes. Cutter replacement indefinitely deferrable.

Impact estimate

Impact estimate — to be validated with your average yield and its dispersion.

The block below is an estimate to be validated against your plant's actual data. We put it forward so the committee has an order of magnitude; we refine it during the assessment.

• Estimated payback 5-10 months.
• Avoided machine-replacement CAPEX: the cutter keeps working and the data comes out from outside. The replacement budget can go to something the plant actually lacks.
• Reduced fabric waste through the parameter-to-waste correlation that cannot be made today.
• Root-cause yield analysis per order goes from impossible to minutes. And the conversation with the fabric supplier about width or shrinkage starts from data, not from impressions.

estimated 5-10 month payback, avoided machine-replacement CAPEX, reduced fabric waste through parameter-to-waste correlation. *Estimate to validate*.

And the fair question from the production manager

“Can a camera reading a screen be trusted for yield analysis?” — here the task is anchored: fixed panel, fields in known positions and physically expected ranges, where the best models drop below 1.5% error [1]. And any reading that falls outside the possible range for speed, pressure or lay height is discarded instead of accepted, so it never pollutes the cutting history. A doubtful reading is marked as such, never silently filled in.

^[1] OpenAI paper "Why Language Models Hallucinate", 2025 — on the reliability of AI in anchored tasks.