AI for Apparel Manufacturing

The Nervous System for
Fashion Supply Chains.

Fabric is 50 to 70% of what a garment costs, and most of its water, dye and energy footprint. We build AI that cuts the waste.

Sheath-Tech builds AI and GPU-native software that cuts fabric waste in apparel, from the cost of a style before it's ordered to the last meter on the cutting table. Three products built. 18 years of domain expertise.

Explore MarkerMind Explore Axiom Explore StyleSync Request a Demo

Recognized by

Member of the NVIDIA Inception program

DPIIT-recognized startup
Startup India

Recognized by
UP Startup Mission

$300B+

Fabric in traded garments every year

98%

Accuracy* vs actual

*Axiom, against actual production consumption, from 100 historical styles

2 to 6%

Of fabric lost at the cutting table. MarkerMind goes after it.

3 Products

MarkerMind in production

Potential environmental impact*

MarkerMind goes after the 2 to 6% lost at the cutting table. Even the low end, 2% saved across the industry, is fabric never spun, dyed or finished.

about 1 million

tonnes of fabric a year

7.4 trillion

liters of water a year

24 million

tonnes of CO2e a year

*At industry scale, 2% of fabric saved. Ellen MacArthur Foundation (2017) clothing fiber volume times WRAP (2017) footprint factors. Fashion is 2 to 8% of global greenhouse gas emissions (UNEP, 2023).

MarkerMind

The Fabric Optimization Engine

  • → Marker optimization. A tighter marker comes from a bigger search. Thousands of GPU cores test placements at once, so MarkerMind reaches markers 1 to 2% tighter than yours, in minutes, for every size ratio and width.
  • → Cut-plan optimization. Every cut plan ranked and fully costed: fabric, spreading, cutting and waste. You pick the plan that fits the floor today.
  • → Fabric end-cut control. A cascading cut plan lets big markers take the bulk and smaller ones consume the end-cuts, with no change to the cutting floor.
Learn More →
Axiom Logo

Axiom

The Operating System for Fashion Costing

  • → 98% accurate against actual production consumption, from 100 historical styles, so brands and factories work from the same number
  • → Parses tech packs and garment images using AI vision models to predict pattern geometry
  • → Deployable in days, not months.
Learn More →

Why Sheath-Tech

The Fabric Trifecta

Fabric is lost in three places: marker making, cut planning and roll end-cuts. MarkerMind plans all three together, so a gain in one is not given back in another.

Built by the Industry, for the Industry

Our founder has spent 18 years in apparel manufacturing and technology, building the tools factories plan fabric with. 500+ factories across 15 countries.

Works With What You Have

Your CAD files in, markers out in the same format. Works as a drop-in solution alongside your existing tools. Hybrid cloud and on-premise. Native ERP/MIS integration from day one.

Product

MarkerMind: Fabric Optimization for Garment Factories

Marker optimization, cut-plan optimization and fabric end-cut control on one GPU engine, under real production constraints. Your CAD files in, markers out in the same format.

Fabric is 50 to 70% of garment cost, and the waste hides in three places: markers that do not pack tight enough, cut plans that are guessed rather than truly optimized, and the end-cuts left over when markers meet real rolls.

These three pull against each other. Most tools fix one and trade off the other two, so the saving on one is canceled out on the others. MarkerMind solves all three together, on one GPU engine.

On a typical order, that recovers 2 to 6% of fabric cost, on top of what your existing tools already save.

Nothing about your setup changes: MarkerMind reads and writes your CAD files natively, whatever system you run, and the cutting floor keeps working the way it already does. Setup takes days, not months.

Three losses. One solve.

MarkerMind optimizes the marker, the cut plan and the roll in one run. Each result shapes the other two, which is why they cannot be solved one at a time.

Marker

Marker optimization

Marker efficiency is a search problem: the more placements a nester can test, the tighter the marker. CPU nesters test them one at a time, so they stop improving after a few minutes and leave fabric on the table. MarkerMind tests them across thousands of GPU cores at once. That is how it reaches a marker 1 to 2% tighter than yours, in minutes, for every size ratio and width in the order.

Cut plan

Cut-plan optimization

Which mix of sizes uses the least fabric, and what is the cheapest way to cut the order? Different size mixes use fabric very differently, and there are too many to check by hand, so most tools guess or test a few. MarkerMind tries every cutting layout, then ranks every plan fully costed: fabric, spreading, cutting and waste. You pick the one that fits the job.

End-cuts

Fabric end-cut control

Cutting from real rolls always leaves something at the end, and rolls do not divide evenly, so the leftovers add up. Smaller markers use them up but waste fabric another way, so most software gave up. A cascading cut plan lets big, efficient markers take the bulk and smaller markers eat the leftovers: near-zero end-cut waste, with no change to your cutting floor.

See It in Action

If you would rather watch than read: GPU nesting on a real order, end-cut planning in action, and MarkerMind's Workbench putting a cutting room's own rules on screen.

GPU nesting on a real order

2 min

End-cut planning in action

2 min

MarkerMind's Workbench

Starts with grouped (married) pieces and shows exactly what each rule costs on a real marker.

Results

Metric Typical today With MarkerMind
Layout generation time Hours per order Minutes
Fabric per order Your current plan 2 to 6% less, on top of your current tools
Marker optimization 1 to 2% lost in the marker Nested on GPU, every marker ranked
Cut-plan optimization 1 to 2% lost in the cut plan Every plan costed, you pick
Fabric end-cut control 2 to 4% lost in end-cuts Near-zero end-cut waste

Who It's For

Apparel manufacturers, cut-room managers, fabric planning teams, and production heads who want to reduce fabric waste without overhauling existing systems. MarkerMind drops in alongside your current setup: your CAD files in, markers out in the same format. It shows whether a better cut plan exists.

Product
Axiom Logo

Axiom: The Operating System for Fashion Costing

98% accurate against actual production consumption, from 100 historical styles, so brands and factories work from the same number. Know what a garment costs before you cut a single piece of fabric.

The Problem

Fabric is the largest line in any garment cost. But at the costing stage, consumption is an estimate, often off by 5 to 10%. For manufacturers, wrong costing is nearly impossible to recoup after price lock-in with brands. The margin is lost before production even begins.

How It Works

Step 1

Upload

Feed Axiom your tech packs, garment images, or specification sheets. AI vision models extract measurements, construction details, and fabric specifications automatically.

Step 2

Predict

AI predicts fabric consumption at 98% against actual production consumption, from 100 historical styles. Human-in-the-loop validation ensures every estimate is reviewed and trusted before it goes forward.

Step 3

Cost

Get evidence-backed fabric cost estimates at the design stage, before a single sample is cut. Turn costing from guesswork into data-driven decisions.

The Small Data Advantage

Axiom reaches production accuracy with about 100 historical styles, so it works from the data a factory or brand already has. Deployable in days, not months.

$300B+

Fabric in traded garments every year

$15 to 30B

Fabric committed on costing estimates every year (5 to 10%)

Sources: Euromonitor (apparel sales 2026); WTO (apparel exports 2024). Ranges from Sheath-Tech field work across 500+ factories.

Who It's For

Costing merchandisers, sourcing heads, and brand product teams who want to move from gut-feel fabric estimates to evidence-backed consumption data. Axiom transforms the costing function from data entry to sourcing intelligence.

Integrations

CAD Compatibility

  • → Your CAD files in, markers out in the same format
  • → Reads standard industry pattern file formats
  • → Drop-in solution: no migration, no re-training
  • → Works alongside your existing tools from day one

Deployment

  • → Hybrid cloud and on-premise deployment
  • → Native ERP/MIS integration from day one
  • → Works alongside your current CAD software
  • → No migrations required. Plugs into your existing stack.
Consumer product

StyleSync: AI Styling Assistant

Upload what you own. Get outfit ideas that actually work. Powered by visual vector search and image generation models.

For You

Upload your closet, get outfit ideas that work. Explore trends. Check if a purchase goes with what you own.

For Stylists & Designers

Generate mood boards in minutes. Explore trend directions. Show clients styled options before they commit.

For Fashion Businesses

Offer virtual styling at a fraction of typical costs. Upload showrooms, build looks, send buyers a link.

Request StyleSync Demo

Powered by visual vector search and image generation models

Reading on fabric, waste and factories

A short list we keep for factory owners, sourcing teams and anyone who wants the practical view of sustainability in apparel.

From our founder

5 reasons why CPL / bit wash is robbing you blind

Cut-panel laundry for knit shrinkage control looks harmless on the floor. Shorter markers, longer end loss and re-planning quietly take 3 to 10% of marker efficiency.

More from Saurav on LinkedIn ↗

Industry research

A New Textiles Economy

The Ellen MacArthur Foundation's baseline for the industry: less than 1% of clothing material is recycled into new clothing, and about 12% of fiber is lost in production.

About Sheath Technologies

Sheath-Tech builds AI and GPU-native software that cuts fabric waste in apparel, from the cost of a style before it's ordered to the last meter on the cutting table. Fabric is 50 to 70% of what a garment costs, and most of its water, dye and energy footprint. Founded in 2020, we bring 18 years of factory-floor experience and applied AI research to an industry that still relies on manual processes and legacy software.

Our products cover the full fabric lifecycle: MarkerMind optimizes how fabric is cut on the factory floor today. Axiom takes that evidence upstream to predict what a garment will need at the design stage. StyleSync brings AI-powered styling to consumers and businesses. Together, they serve the value chain from design to sourcing to production.

Sheath Technologies Private Limited, verified on Tracxn ↗

Founded 2020

India

3 Products

MarkerMind, Axiom, StyleSync

B2B + B2C

Enterprise AI + Consumer AI

18 Years

Domain expertise in apparel

Founder

Founded by Saurav Ujjain, who has spent 18 years in the fashion supply chain, from factory floors to enterprise technology to AI, and has been full-time on Sheath-Tech since January 2025. He holds a BFTech in Apparel Manufacturing from NIFT Delhi, an Executive PG in AI/ML from IIIT Bangalore and an MSc in AI/ML from LJMU (2026). He has led fashion-tech businesses and scaled a factory software platform across 7 countries. He has worked with 500+ factories across 15 countries, won a $400K AI transformation challenge, and built Sheath-Tech's three products: MarkerMind, Axiom and StyleSync.

LinkedIn ↗

Get in Touch

Tell us about your requirements. We'll get back to you within 24 hours.

We respect your privacy. No spam, ever. Your message is relayed to us by FormSubmit.