No file loaded
Drop a Tajima .dst file into the left panel. Analysis runs entirely in your browser — files never leave your device.
Design Preview
SVG will appear here after analysis
| # | Start (mm) | End (mm) | Len (mm) | Type | Color |
|---|
Threshold Sweep
Count of segments exceeding each length threshold, broken down by type
| Threshold | Normal stitches | Jump (trim) | Jump (travel) |
|---|
Ranked Findings
Numeric-impact findings first (biggest estimated runtime/time cost), then the rest by severity · 0 findings
Stitch-Type Classification
Non-overlapping window turn-angle / orientation-concentration classification, weighted by stitch count · substrate: twill
| Type | Stitches | % of total | Blocks |
|---|
Low-Confidence Blocks
Classified blocks below 0.70 confidence — inspect individually rather than trusting the aggregate
| Block | Type | Confidence | Stitches |
|---|
Stitch Length
Per-block hard-minimum violations (critical) and safe-zone stitches (warning, sewable but risky) · worst first
| Block | Type | Stitches | Below Hard Min | Safe Zone | Over Long |
|---|
Density
Satin column spacing/width (PCA advance axis) and fill row spacing, per block
Satin columns
| Block | Stitches | Spacing Mean | Spacing Median | Width Mean | Flag |
|---|
Fill rows
| Block | Rows | Mean Row Spacing | Flag |
|---|
Fill Angle
Dominant stitch direction histogram across all fill blocks (0–180°, mod-180 — direction is ambiguous within a row)
Layering
Cross-object penetration analysis, scoped to the whole file across all colors and stitch blocks — penetration density (§5a), near-coincident penetrations (§5b), thread-length-per-cell (§5c)
Penetration density heatmap (§5a)
0
0
Speed & Calibration (§6)
Per-stitch achievable-SPM model — machine ceiling, long-stitch ramp, long-move window spread, short-stitch ramp, and wide-satin ramp, combined by MINIMUM (the slowest factor wins) · plus flat trim/color-change overhead. Driven by the sidebar's Runtime Estimate fields and the Substrate selector.
Binding speed factor — which limiter set the achievable SPM, by stitch count
| Factor | Stitches |
|---|
Estimated time lost to each penalty — marginal cost, if only that one factor were removed
| Penalty | Stitches | ~Seconds |
|---|
Calibrate against an observed stitch-out
A simple heuristic, not an exact solve — closing a predicted/actual gap could mean the
machine speed ceiling is off, or that trim/color-change overhead is under- or over-counted.
Which is the better explanation is left to your judgment (e.g. a gap that scales with
stitch-dense vs. stitch-sparse designs points to the SPM side; one that's roughly constant
across very different designs points to overhead).
Actual stitch-out time (minutes)
Limitations
What this tool can't tell from the DST file alone:
Compare: vs
Key deltas (B minus A; negative = improvement for time/violations) · substrate:
Side-by-Side Metrics
Every scalar metric the report produces, both files side by side
| Metric | File A | File B | Delta (B−A) |
|---|