HomePlatformSolutionsAnalyticsPricingCompany
Try the demo →
Diagnostics & metrics

The metrics, exactly as your team sees them

These are the live dashboard widgets — running on sample data here so you can understand what each one tells you before you ever sign in.

Open the full demo →See the workflows
Quality gallery

Four-point grade, SPC control, Pareto, quality index & more

The same diagnostics bento that sits on every dashboard and production line — industry-standard QC metrics computed live from inspection data.

Diagnostics & Metrics

Quality Index

DHU

Defects per hundred units · live grading

216.3DHU
Reject
0260
Acceptance limit≤ 15 DHU
Top contributors
Needle mark
31.0
Stain
29.6
Hole
29.1

Fabric Grade

4-POINT

Penalty points per 100 units · ASTM-style

Reject
460pts/100u
ABCReject
Grade-A limit40
934 points203 units

Defect Pareto

Top 6 types drive ~80% of defects

Process Control

12 out of control

Defects / unit · u-chart with ±3σ limits

Defect Heatmap
439 · 203img
top-right · 53%
Low
High
Needle mark63·Stain60·Hole59
0
Severity Index
Elevated
0–100
Density
2.2/img
Confidence
91%
7-day
68 today

Supplier Scorecard

Suppliers ranked by quality

1
Anatolia Textiles26.1
2
Izmir Indigo31.3
3
Acme Dye Supply108.3
4
Bursa Fabric Co.163.9
5
Marmara Colour283.3
6
Hassan Tex603.8

Cost of Poor Quality

Projected loss from defective units

$3.0kest. loss
$15/inspected unit · 103 defective of 203
$2.1k scrap$927 rework

Run Drift

→ Stable

Defect density across the production run

Run startLatest
Attribution & forecast

Who, which machine, which shift — and what’s coming

Slice defects per 100 units by machine, operator or shift, and get a forward forecast of when a line will breach its control limit.

Defects by Machine

Defects per 100 units · highest first

Loom L-5612.9 /100u · 190 in 31
Loom L-3590.5 /100u · 124 in 21
Loom L-7163.9 /100u · 59 in 36
Jet Dyer D-2104.3 /100u · 48 in 46
Finisher F-126.1 /100u · 18 in 69

Early Warning

low confidence

Forecast to UCL · Loom L-3 · from run drift

Trend stable
Now 5.40 /unit · UCL 8.32 /unit
Trend -4.43 /100uCenter 6.04 /unit
Model health & roll map

Confusion matrix & meter-marked defects

Model-vs-human agreement on the left; the physical roll defect map for cut planning on the right.

Model vs. Human

Confusion matrix from 231 reviewed detections

Agreement
91.9%
Precision
87.3%
Recall
87.3%
pred ╲ actualHoleStainStreakWeftWarpNeedleBrokenPinchedKnotNone
Hole583·······2
Stain241·······3
Streak··24·4····6
Weft··319······
Warp··2·17·····
Needle·····11··2·
Broken······9·1·
Pinched·······6··
Knot········7·
None1····5·32·
HoleP 92.1% · R 95.1% · F1 93.5%
StainP 89.1% · R 93.2% · F1 91.1%
StreakP 70.6% · R 82.8% · F1 76.2%
Weft defectP 86.4% · R 100.0% · F1 92.7%
Warp defectP 89.5% · R 81.0% · F1 85.0%
Needle markP 84.6% · R 68.8% · F1 75.9%
Broken stitchP 90.0% · R 100.0% · F1 94.7%
Pinched fabricP 100.0% · R 66.7% · F1 80.0%
Knot / SlubP 100.0% · R 58.3% · F1 73.7%

Roll Defect Map

136 m · 84 defects

Defect position along the roll · plan cuts around the marks

0m27m54m82m109m136m
Longest clear cut 16.1 m at 67 mDefects / 10 m 6.2

Run these on your own data

Sign in to your workspace, or explore everything with sample data first.

Try the live demo →
Talk to sales
Tell us about your production line. We'll reach out within 24 hours.
We read every message. Responses typically within one business day.